With nine weeks until the midterms, here’s the bottom line from our team at Protect Democracy: everyone can and should participate in our democracy this fall with confidence.

That’s not to ignore how acute and alarming the dangers are. But we have an extensive toolkit, a track record to back it up, and our ecosystem is prepared and well-coordinated. We’ve survived assaults on our elections before and we’re going to do it again.

Subscribe now

When read individually, the eight threats below can feel overwhelming. But when we understand them as part of the Deceive, Disrupt, Deny playbook, we can channel our energy into deploying a stronger playbook of our own. That’s why our team is partnering with organizations across the ideological spectrum and key sectors — labor, faith, business, public health, Gen Z, national security — to prepare their members to respond to each of the threats below. We’re also utilizing our own communications strategy for reaching and educating voters, including on the digital platforms where most Americans now get their news, so that all of us recognize the election interference playbook and understand how we can respond.

Read on for a closer look at the scenarios we’re preparing for and why we’re confident that our movement will rise to the challenge. Of course, this isn’t everything that our 125+ experts are working on — and it’s a tiny portion of what all of us, the pro-democracy majority in this country, are doing together.

Threat one: The Justice Department may attempt to seize ballots and voting machines

Starting in January, when the FBI relied on baseless claims of fraud in 2020 to raid an election center in Fulton County, Georgia, the Trump administration has laid the groundwork to try to disrupt the election by seizing materials or otherwise messing with election administration.

We’ve seen equally worrisome investigative actions in Arizona, Michigan, Wisconsin, and California. As our friends at Just Security put it: “For the first time, the country faces the possibility that the DOJ may attempt to take over polling places, seize ballots during an election, or threaten to prosecute voters or election officials to interfere in the administration of an election.”

How we’re responding: Since an ounce of prevention is worth a pound of cure, Protect Democracy and our peers are producing materials for the judges who evaluate these seizure warrants to raise the odds that they reject deficient and politically motivated applications.

And if seizures do happen, we’re making sure election officials have the protocols and resources in place to minimize harm to election operations and vote counting. With the Brennan Center, we’ve published and widely distributed guidance on lining up counsel before the pressure hits, maintaining redundancies of records, documenting everything if federal officers show up, and being prepared to file for emergency relief. This includes state-specific guides we’re producing to help officials create personalized plans of action.

We’re also coordinating with national peers as well as state and local organizations to ensure that, if election officials or other actors need to file emergency litigation to halt blatantly illegal attempts to seize election materials or interfere with election administration, they have the support to do so rapidly and forcefully.

Threat two: The president is actively seeking to seize state voter files and impose barriers to the ballot

The president is pursuing a sweeping campaign to try to seize control of who gets to vote and how. Trump has issued executive orders requiring documentary proof of citizenship on federal registration forms and directing DHS to dictate to USPS which voters can receive a mail ballot. The Justice Department has demanded that every state hand over its complete voter rolls — data which could be used to force purges based on bad matches and manufacture doubt about results. DHS and FEMA are attempting to condition billions in counterterrorism funding on states turning over that same data. DHS is also invading the privacy of millions of Americans to assemble and compile a massive, inaccurate master “citizen list” which it now plans to pressure election officials to compare against state voter rolls. And the administration has renewed its push for the SAVE America Act, which would require documentary proof of citizenship nationwide (but which fortunately does not have the votes to pass the Senate).

How we’re responding: Our ecosystem is making good progress already. First, pro-democracy actors have blocked core provisions of both of Trump’s anti-voting executive orders (and as we explained here, the Supreme Court USPS ruling is almost certainly noise, not substance). Second, the Justice Department is 0-23 in its lawsuits against states that have refused to hand over their voter rolls; Protect Democracy also filed a lawsuit against the Justice Department with co-counsel at the ACLU, ACLU-D.C., Citizens for Responsibility and Ethics in Washington, and the Democracy and Rule of Law Clinic at Harvard Law School to block the Justice Department’s illegal attempts to compile a national voter database. Third, a coalition of state Attorneys General sued to block DHS and FEMA from conditioning counterterrorism funding on states adopting federally dictated election administration conditions; and with the Jacobson Lawyers Group, Brennan Center, and Public Rights Project, Protect Democracy just filed a related lawsuit against DHS to ensure the relief applies to as many jurisdictions as possible. Fourth, along with Citizens for Responsibility and Ethics in Washington and The Electronic Privacy Information Center, Protect Democracy just filed a lawsuit to prevent the administration from pressuring states to make last-minute changes to their voter rolls based on DHS’ erroneous master “citizen list.”

Threat three: ICE/DHS deployments may seek to intimidate or suppress voters

Administration actors and allies have long hinted at DHS/ICE presence during the midterms. Steve Bannon called ICE deployment to airports a “test run.” Former White House Press Secretary Karoline Leavitt said that she “can’t guarantee that an ICE agent won’t be around a polling location in November,” and Attorney General Todd Blanche publicly asked, “Why is there objection to sending ICE officers to polling places?”

Blanche later refused to explicitly rule out deployments at poll sites. Meanwhile, when the DNC sued for records of any plans to deploy agents to polls, ICE told a judge no records exist — and then reversed course and admitted to having more than 11,000 potentially relevant pages; CBP found nearly 117,000. Bottom line, we have to prepare for ICE/DHS to be used as a tool for election interference, whether through deployment at polling places, or more likely, through surges of federal presence in competitive districts under the pretext of routine immigration enforcement.

How we’re responding: Here, both litigation and legislation are potentially effective.

On the legal front, we’re preparing for emergency litigation if federal agents are deployed in ways that intimidate voters or suppress participation. We are ready to seek emergency relief and can build on our work in Arizona in 2022, when we persuaded a Trump-appointed judge to strictly limit armed vigilantes’ behavior at ballot drop boxes.

Meanwhile, on the legislative front, five states have passed legislation based on a model bill that our team crafted to let people sue federal officials who violate their constitutional rights — with California the most recent to cross the finish line on August 27th. Once the California bill is signed by the Governor, more than 70 million Americans will live in a state with this kind of legal protection. As voting begins, these laws can be used to challenge actions that deprive people of the right to vote, including intimidation and violence at poll sites. Agents who know their actions carry personal liability are also less likely to commit abuses.

Threat four: Surveillance and policing by ICE and DHS may put a chill on First Amendment activity

Surveillance, intimidation, and overzealous law enforcement have the potential to chill Americans’ exercise of their rights to protest and observe government action.

Consider a growing trend: an investigation by ProPublica and PBS identified more than 300 people arrested at anti-ICE protests and charged with crimes like assaulting or interfering with federal agents — and found that more than a third of those cases have collapsed, either through acquittals or dropped charges once officers’ claims didn’t hold up.

Our team spoke to NPR in a similar piece about the free speech and First Amendment implications of how the Justice Department is weaponizing criminal conspiracy charges against protesters. In addition, The Wall Street Journal has reported on how ICE has built a $258 million surveillance operation scanning over 200 websites for critical speech and issuing hundreds of subpoenas to de-anonymize posters.

How we’re responding: We’re working to rein in the government’s surveillance and intimidation tactics through litigation and legal advocacy.

In February, we sued DHS on behalf of Maine residents who were told they were being added to “domestic terrorist” watchlists for lawfully filming ICE operations.

At a March hearing, the government’s lawyer couldn’t tell the judge what the government is doing with our clients’ personal information, including where it’s being kept and how it’s being used. As Time Magazine wrote, the case gets at a question “never supposed to be this difficult to answer: how is the government treating ordinary citizens with whom it disagrees?”

Meanwhile, we just filed a brief from prominent First Amendment and rule of law organizations — including the Foundation for Individual Rights and Expression, Society for the Rule of Law, and the Cato Institute — in a criminal conspiracy case against Minnesota activists. The brief urges the court to order the government to turn over documents that could demonstrate political motivation behind the prosecution; it also criticizes the government’s surveillance tactics and argues that the scale of this assault on free expression has “not been seen since the Second Red Scare.”

As the government’s data-gathering against critics and observers expands, this work can help keep it in check. And in addition to the privacy of our data, we’re also working to uphold the privacy and sanctity of our homes. In April, we brought a lawsuit against DHS to strike down the agency’s warrantless home entry policy, under which armed, masked agents are entering and searching homes without a judge’s permission.

In each of these cases, we can build precedent to constrain lawless behavior and deter similar election-related overreach.

Threat five: Government sources are working to corrupt the information ecosystem and seed propaganda

This is two threats in one. Our colleague Rachel Goodman summed up the first:

Trump FCC Chair Brendan Carr has wielded “public interest” to threaten and investigate broadcasters for speech as core to the First Amendment as editing an interview with then-presidential candidate Kamala Harris; reporting on local ICE operations; covering the Iran war in ways that the President dislikes; and continuing to air Jimmy Kimmel’s comedy in the face of White House opposition… And Carr’s threats run not only to broadcast licensees, but to the media companies that own and operate them…

The second threat centers on whether the federal government has used public money to produce and distribute favorable content about its own immigration operations, and whether it has granted access to commentators who cover it favorably while excluding independent journalists. For instance, in Portland, Oregon, the government has not disputed deploying chemical munitions in a way connected to creating content, including staged video, to support a false narrative about the city. Elsewhere, agencies have invited favorable commentators on immigration ride-alongs and into facilities while turning away reporters who requested the same access.

How we’re responding: On the first front, in July, we brought a filing from a bipartisan coalition of former FCC officials opposing FCC Chair Brendan Carr’s latest effort to revoke the licenses of broadcasters whose speech the Trump administration dislikes — in this case, ABC. We brought a similar filing in 2025 in connection with the FCC’s highly unusual proceedings against CBS. Our latest filing received widespread coverage, including from unlikely validators like Newsmax.

We also recently won a preliminary injunction in our lawsuit against a Trump administration immigration policy that targets tech researchers, content moderators, and trust-and-safety professionals whose work counters election misinformation.

On the second front, we’re working to force the release of the information we need to know whether the federal government used public money or special access to create favorable coverage of its immigration operations. As our colleague Katie Schwartzmann put it:

“Americans have a right to know whether our tax dollars are funding a propaganda effort — one that distorts the news, pumps out flattering coverage, and freezes out reporters who would scrutinize the Administration’s narratives.”

In response, Protect Democracy filed Freedom of Information Act requests seeking relevant records from ICE, CBP, and DHS. The government failed to comply with our requests by the deadline set by law, so we brought a lawsuit asking a court to compel disclosure of the requested records.

Threat six: State officials are likely to face pressure to subvert the certification process

Election certification is a mandatory duty for the officials involved — and one that they have no option to refuse. But local and state officials are increasingly facing intense pressure to withhold their signatures until conspiracy theories are “investigated” or results are overturned.

There’s the infamous example from 2020, when President Trump asked Georgia Secretary of State Brad Raffensperger to “find 11,780 votes.” Days earlier, Trump had personally pressured two Wayne County, Michigan canvassers not to certify results showing he’d lost the county. By our count, at least 21 counties in eight states have seen attempts to refuse to certify since 2020; we know that the certification process is a likely target, especially since many of the officials who have upheld their oath in prior cycles have been replaced with more pliable actors.

How we’re responding: We’ve built and refined an effective certification protection strategy in past cycles. Our in-state advocates have built strong bipartisan working relationships with election officials in key states, which allow us to share resources like these explainers with these actors to reaffirm their constitutional requirements.

Recognizing that these officials are facing tremendous pressure, we’ve also mobilized public and private support — for instance, by running radio and TV campaigns in their districts explaining that certification is mandatory, and by organizing outreach from influential community members like business executives and faith leaders to provide them with direct encouragement to hold the line.

Lastly, if needed, we can apply counterpressure, for instance by ensuring that state officials are prepared to seek court orders to force local bodies to certify, and by contributing to media coverage that deters officials from abuses by reminding them that they could face criminal liability for failing to follow the process.

Threat seven: Litigation may seek to change the rules of the game and/or flip the results

Election deniers have developed a specific playbook for this: the “zombie lawsuit.”

These are cases filed prior to the election, not with the goal of winning beforehand, but instead with the intention of planting a complaint that “rises from the dead” so that a court or Congress has a pretext to overturn results if the litigants’ preferred candidate loses.

In 2024, we tracked dozens of these across swing states, including suits attempting to disqualify military and overseas voters’ ballots in Pennsylvania, Michigan, and North Carolina, and challenges to routine voter list maintenance in Nevada, Arizona, and Wisconsin. All of them failed, but we can nevertheless expect to see these kinds of claims rise from the dead (again) this fall.

How we’re responding: The good news is that we’ve defeated this threat before. Consider the most serious of the post-2024 zombie lawsuits: in North Carolina, Justice Alison Riggs won her Supreme Court race against Judge Jefferson Griffin by 734 votes, which was confirmed by two recounts. Griffin then tried to throw out roughly 65,000 votes and the state Supreme Court initially sided with him.

We used a broad toolkit in response. Our in-house data experts helped identify voters who lawfully cast their ballots and were at risk, including many who voted for Griffin. We worked with partner organizations to place these voters’ stories in coverage across the state. We also helped organize a rebuke from 200+ former Republican elected officials and conservative members of the bench and bar. Lastly, after the state Supreme Court issued its opinion in Griffin’s favor, we filed a class-action lawsuit of impacted voters in federal court.

In May 2025, a Trump-appointed federal judge ordered the state Board of Elections to certify the actual result. Griffin then conceded the race. We’re prepared for this kind of post-election litigation in key states this cycle, building on what we learned from this win.

Threat eight: Members of Congress could try to refuse to seat the duly elected winners in January

As Molly Roberts and Lawfare explained here, the process for seating members of Congress could become a target for bad actors seeking to override the rightful results.

It’s a remote possibility, but the stakes are high, so we have been preparing for scenarios where duly elected candidates could be denied their seats. This could play out in a number of ways: bad actors trying to manipulate the roll of members allowed onto the floor during the seating process; members-elect exploiting a tradition that allows any member to object to another’s swearing-in; and state-level certification delays or false certificates that muddy who the “rightful” winner is. As an example, Senator Tommy Tuberville has already previewed the kind of baseless claims that could fuel these tactics, saying without evidence that “We probably have four or five senators that didn’t legally win. They shouldn’t be up here.”

How we’re responding: It’s first worth noting how exceptionally difficult it would be for bad actors to pull this off. As Prof. Derek Muller explains on the Election Law Blog:

In short, most scenarios get basic facts or law wrong, and the most complicated scenarios require a unique convergence of factors that have never existed, a unique level of coordination that seems very unlikely, docile political opposition, and a judiciary willing to acquiesce to all these moves–and all working simultaneously. Even then, most of these scenarios could, at most, delay some organization of the House by a matter of hours or days.

That said, there are congressional, legal, and public readiness strategies we’re deploying. We’re in the weeds of the state-by-state and procedural vulnerabilities and solutions, including working with right-of-center partners to ensure that enough House Republicans are ready and willing to oppose any attempt to manipulate seating.

We’re also working with partners to train congressional staff on vulnerabilities and procedural steps for resolving election disputes. And, as reported by NPR, Politico, and The Guardian, we’re helping key members, including Democratic leadership, prepare decisive responses to various potential crises well before any of them could play out.

Lastly, on the public readiness front, we’re helping assemble the broad, cross-sector relationships to mobilize a rapid response should there be a credible attempt to overturn the results.

Bonus threat: President Trump could invoke emergency powers, the Insurrection Act, or some other pretext to interfere with voting and election administration

Trump has several tools he could reach for — he could declare a national emergency, which unlocks a sprawling set of statutory powers; he could invoke the Insurrection Act, which allows military deployment for domestic law enforcement; or he could try to federalize the National Guard through other statutory authority. Even a legally doomed attempt at using any of these could do damage before a court steps in, by creating chaos and confusion, intimidating voters, and handing Trump a story to tell about how he tried to secure elections and the courts undermined him.

How we’re responding: The good news is that as scary as they sound, these scenarios are a bit of a red herring.

None of these paths actually get Trump what he would need to interfere with voting and election administration. Congress gave control of elections to the states, and no emergency declaration or statutory invocation changes that — or converts otherwise illegal actions into legitimate ones.

We’ve already seen how this plays out in practice: throughout 2025 and into 2026, courts blocked Trump’s attempts to federalize and deploy the National Guard into several cities. (Protect Democracy filed briefs repeatedly on behalf of former senior military leaders explaining that these deployments are dangerous and illegal.) In short, this is mostly about optics and political theater; in 2025, Trump declared more national emergencies in his first 100 days than any president managed in a full year, but this hasn’t changed who actually runs elections or how courts have ruled when he’s tried to stretch these authorities past their limits.

Combine that with his public unpopularity and polling data that shows domestic deployments are deeply unpopular, and any attempt along these lines seems more likely to backfire than to provide him any meaningful gains.

So what does this all mean for November?

The threats are real and alarming, but we should be confident in our collective ability to participate freely and fairly in our democracy this fall.

That word — participate — is the key one. Collective action is the only way out. Thousands, if not millions, of Americans are hard at work in ways big and small to defend the midterms.

We all have a role to play in upholding the foundations of our democracy.

If you’re wondering how to get involved, what more you can be doing personally, and which actions you can encourage your friends and families to take, we’ve got you covered. This playbook is updated weekly: How you can protect democracy.

We all need to do our part — not just by taking all the actions on this page, but also by sharing them with our communities.

Together, we can ensure that our movement for freedom and democracy prevails.

A short, mid-week explainer of an important lawsuit my colleagues filed today.

Beyond the USPS mess (which I wrote about last Thursday), there’s another part of the March executive order on elections that’s arguably just as important: The federal government is now attempting to build and distribute a massive database of personal information on hundreds of millions of Americans.

Subscribe now

As I write this, the Department of Homeland Security, the US Citizenship and Immigration Services (USCIS), the Social Security Administration, and the State Department are working to compile all of our information into one massive federal database that they plan to distribute back to states. All of our information — including potentially sensitive data, like Social Security numbers — would be consolidated in one place.

The goal, of course, is to create election chaos. How?

The Trump administration plans to send all this compiled data to election officials in all 50 states in the form of “State Citizenship Lists” and attempt to have them use those lists for the November election. “Not on the list? Well, sorry. Take it up with DHS.”

Like with USPS, it’s a delusional plan and extremely unlikely to work (for one, it’s very illegal — see below). But you can probably see all the different ways even a good-faith version of this database would be a disaster:

• Election chaos like you can’t imagine: Coordinating individual data across all 50 states is a herculean effort. It took about 20 years for state and federal governments to roll out REAL ID. And here the Trump administration wants states to implement a completely new voter roll process — “here, review, edit, and implement this list of eligible voters based on our garbage data” — barely 60 days before Election Day. It would make the DOGE chaos look like a well-oiled machine in comparison.

• Errors! Errors! Horrifyingly, DHS has admitted that the data that it intends to use is full of inaccuracies that could wrongly flag eligible voters as ineligible to vote. Google your own name, right now. How many other people have the same first and last name as you? What if one of them isn’t an eligible voter and DHS gets the two of you mixed up? (This sort of record-keeping is much harder than it seems.) Or did you get married in recent years? Move addresses? Are you a recently naturalized citizen? There are dozens, if not hundreds, of different ways data errors in this database could lead to lawful citizens losing their constitutional right to vote.

• A whole-country data privacy vulnerability: There are something like 240 million citizens over 18 in the United States. I don’t think it needs to be said that putting all of our information — including potentially Social Security numbers — into one big database and then distributing that information to the more than 10,000 election jurisdictions across the whole country is a very, very, very bad idea.

• Federal control over elections: This database would give the federal government unprecedented power over elections. A list of eligible citizens may seem like a simple thing — after all, noncitizen voting is already illegal and, by every credible measure, basically nonexistent — but the list itself would be a powerful tool that could be weaponized against states. What happens if the administration attempts to overrule state voting rights restorations and exclude whole categories of voters? Or what if this list becomes a vector for the ongoing quest to overturn birthright citizenship? Control of voter lists is control over elections, period.

• A powerful vector to attack election officials: The scariest part of this plan may be what happens if the federal database says you’re not eligible (beyond just you potentially being disenfranchised). DHS has threatened to investigate and prosecute states and election officials who “issue Federal ballots to individuals not eligible to vote.” If the lists disagree — if state data says you’re eligible to vote but DHS says you’re not — they may try to put your election officials in jail for it.1 In other words, this whole plan may end up being a trojan horse to directly attack the people who run free and fair elections.

Last week, DHS claimed it found 15,903 noncitizens on the Nevada voter rolls. Asked for the list, it produced just 185 names — with no explanation of its supposed 15,718 others.2

This plan would not go well.

The master database is as illegal as it would be disastrous

Today, on behalf of the Electronic Privacy Information Center (EPIC), Protect Democracy and Citizens for Responsibility and Ethics in Washington (CREW) sued the Trump administration over this master database.

Said John Davisson, deputy director and director of enforcement at EPIC:

In the United States, we have limits on how government agencies can collect and use sensitive personal data—limits this administration is blatantly ignoring… Building a database containing the identities, social security numbers, addresses, and citizenship information of every U.S. citizen is an extraordinary illegal overreach by the federal government. This is precisely the nightmare ‘Big Brother’ scenario that the Privacy Act exists to prevent. We hope the court will block this effort immediately and protect every American’s fundamental right to privacy.

The lawsuit argues that this plan is illegal in three different ways:

1. It’s an unconstitutional power grab. The Constitution is clear: “The Times, Places and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof; but the Congress may at any time by Law make or alter such Regulations.” You don’t need a law degree to see how there’s nothing in there about the president having any say over who can and can’t vote.

2. This plan violates federal privacy laws. To make these lists, the administration wants to pool Americans’ sensitive personal data — your sensitive personal data, my sensitive personal data — without notice, without consent, and without any way for us to correct errors. Federal privacy laws, including the Privacy Act, were written explicitly to prevent this kind of unauthorized data consolidation and disclosure. The government simply can’t play fast and loose with data like this.

3. This scheme threatens to disenfranchise eligible voters and sow chaos ahead of the midterms. Finally, most practically, it’s illegal in its likely effects. The government has admitted the data behind these lists is riddled with errors. The same flawed data has already led to eligible voters being wrongly flagged as ineligible. With state officials pressured to use these lists just weeks before Election Day, the result could be widespread confusion, wrongful purges, and eligible Americans turned away from the polls. Illegal, illegal, illegal.

Read more about the case here.

Would love help spreading the word. Send this to people in your life who care about privacy. And/or give us a re-post? 👇

The FCC’s broad power to require broadcast media to operate in “the public interest, convenience, and necessity” has long been a loaded gun available for abuse, and it has lately been used to justify a shocking rise in government speech suppression. The next Congress must pass legislation to narrow and define it.

Subscribe now

Trump FCC Chair Brendan Carr has wielded “public interest” to threaten and investigate broadcasters for speech as core to the First Amendment as editing an interview with then-presidential candidate Kamala Harris; reporting on local ICE operations; covering the Iran war in ways that the President dislikes; and continuing to air Jimmy Kimmel’s comedy in the face of White House opposition. Carr recently “reminded” broadcasters of their public interest obligations, in a document that consigns the statutory prohibition on FCC censorship to a footnote and mentions the First Amendment only to discuss the ways in which it arguably does not apply to broadcasters. And Carr’s threats run not only to broadcast licensees, but to the media companies that own and operate them; thus, he has weaponized the public interest authority to leverage licenses of ABC’s owned and operated stations to attack the speech of Jimmy Kimmel and the corporate practices of the Disney corporation.

It comes as no surprise that the second Trump administration has transformed the FCC into a regulatory cudgel aimed at both suppressing opposition speech and rewarding administration-friendly speech; this weaponization of regulatory authority is how modern authoritarians capture journalism. In Hungary, Viktor Orbán ran this playbook perhaps most effectively, establishing a Media Council that wielded its licensing and merger review authority to successfully undermine independent media and ensure that his allies controlled all major media outlets.

The public interest standard governs the awarding, renewal, and transfer of licenses (that is, merger approval). The broad authority that it provides the agency over licensees leaves both broadcasters and their corporate parents exposed to government jawboning. The language dates to 1927, when scarcity of radio frequency was a major problem and the modern conception of free speech had not yet coalesced in the Supreme Court or the public imagination. Today, multivocality exists far beyond the airwaves, and the vestigial public interest standard has been transmuted into a tool of abuse. And, while the recent volume and frequency of abuse is unprecedented, Democrats have not been immune to the temptation to use the authority to suppress speech.

Meanwhile, there’s a growing consensus on the libertarian right that the FCC should be eliminated entirely, in part based on these abuses. But the specific problem that Carr’s term has highlighted is the breadth of the FCC’s content regulation authorities. The Supreme Court could take a major step toward solving this problem by updating its jurisprudence on First Amendment rights in the broadcast context to account for these developments — its 1969 decision in Red Lion is its most recent decision directly on point — but it is Congress that is best situated to solve the problem by trimming the relevant statutory authority.

Congress must replace the FCC’s current broad and vague public interest authority with a definition explicitly enumerating permitted areas of regulation — and with respect to content regulation authorities those should be exceedingly narrow. Reform legislation must make clear that there is no residual, unenumerated FCC power to regulate speech. This specific definition should foreclose the FCC from, for example, sua sponte determining that “partisanship” violates the public interest standard and imposing license consequences based on its assessment of a station’s partisan content.

It is imperative that legislation explicitly repeal the “news distortion” policy, an uncodified authority to which the FCC has laid claim for decades and which has been a key vector of recent abuse. A bipartisan coalition of former FCC commissioners and chairs (which my organization represents) have united behind the need to repeal the policy in order to protect speech. As the Supreme Court recently put it in Moody v. NetChoice, “it is no job for government to decide what counts as the right balance of private expression—to ‘un-bias’ what it thinks biased, rather than to leave such judgments to speakers and their audiences.”

Congress should also consider whether other content regulations, such as children’s television requirements and indecency and obscenity rules, still make sense in today’s media environment — with an eye to the fact that in the context of broadcast licensing, all regulatory requirements risk becoming vectors for government suppression of speech. Current statutes also give the FCC election-related content regulation authorities — the separately codified equal opportunities and lowest unit charge requirements. These each have played crucial roles in the election information ecosystem, and have shaped Americans’ expectations about it, such that removing them abruptly would create real confusion. But they also create opportunities for abuse, and more thinking about long-term reforms is needed.

Simultaneously, Congress should strengthen section 326 of the Communications Act, the anti-censorship provision. That language makes clear that the FCC lacks any “power of censorship… and no regulation or condition shall be promulgated or fixed by the Commission which shall interfere with the right of free speech”. But the FCC and the Supreme Court have suggested the provision mostly concerns prior restraints, and it has not prevented the FCC from imposing license consequences for content after the fact. Reform legislation should clarify that the FCC may never impose a regulation or condition, including any license consequences, for licensee speech beyond the enumerated speech authorities — and any action at all that constitutes viewpoint discrimination or retaliation.

Congress should further strengthen the ban on censorship by including a private right of action along the lines of the one described in the recently introduced JAWBONE Act — so that the burden is on the agency to demonstrate that its action does not violate the statute or the First Amendment.

Finding solutions to protect speech has never been more urgent; the Supreme Court’s decision in Slaughter has called into question the status of the FCC as an independent agency — and Chair Carr has disclaimed any independence. While this shift may necessitate deeper structural changes to the agency going forward, legislative amendments to curtail the public interest authority are the most narrowly tailored solution to the speech suppression problem lately posed by the FCC.

This article was originally published by the Knight First Amendment Institute at Columbia University.

This week, the White House announced that it has finalized one of the more important documents in the country right now.

It’s supposedly a new framework for the Trump administration to review advanced AI models and decide whether and under what conditions to allow those models to be released to consumers and the public.

Subscribe now

Except for a high-level description, though, it did not announce what’s in the document. The administration has privately shared the new framework with a handful of frontier AI companies for their input — but beyond that, the most consequential federal regulatory framework to date in the United States for the most significant technology in the world remains secret.

We don’t know who developed this frontier AI evaluation framework, who influenced it, and how. We don’t know what it contains, how it will be applied, or by whom. And we don’t know the legal authority that purportedly undergirds it. We do know Congress had no role.

This is not how laws are supposed to be made in a democracy. And it’s especially fraught given this administration’s penchant for aggrandizing and abusing its own power to enrich itself, reward its loyalists, and punish those who don’t go along.

So we are investigating.

We just submitted public records demands to the Department of Commerce, Treasury, the White House Office of the National Cyber Director, and the White House Office of Science and Technology Policy. We asked for records on an expedited basis identifying the framework’s terms and criteria, the participating companies, and the legal authority for establishing and operating it. You can read our FOIA requests here.

The White House’s recent actions on AI models

Before we get into why we’re so concerned about this, a bit of background.

There has been extensive debate in Congress and in state legislatures on the many reasons that our government might regulate AI. These include protecting our kids, guarding against disinformation, safeguarding privacy, ensuring economic fairness, strengthening cybersecurity, guarding against catastrophic risk, and beyond. Some states have enacted AI regulatory legislation on these types of issues. In our view, it’s vital for Congress to legislate in ways that maintain democratic control over how AI is developed, deployed, and affects our society. As of now, however, the U.S. Congress has not done so — nor has it authorized the White House to enact its own legal regime governing AI.

Nonetheless, the White House suddenly has sought to exercise extensive regulatory control over aspects of the industry.

After initially taking a hands-off approach to frontier AI in the name of winning the race against China, the Trump administration made a dramatic shift toward pre-release review when the progress of frontier AI began to manifest some of the security risks experts had predicted. On June 2, 2026, the White House issued Executive Order 14409, “Promoting Advanced Artificial Intelligence Innovation and Security,” which claimed to create a “voluntary” pre-release review program for frontier models. In reality, it established what Dean Ball, former Trump administration AI policy advisor now leading frontier AI policy at OpenAI, described as “a de facto involuntary licensing/preapproval regime.” And one that operates behind closed doors pursuant to undisclosed terms and murky legal authority. The administration’s subsequent actions pursuant to the EO proved this to be true.

On June 12, 2026, the Department of Commerce ordered Anthropic to suspend worldwide access to two of its AI models, Fable 5 and Mythos 5, within ninety minutes, citing export-control authority in an unprecedented context. The Commerce Department later restored access to Mythos 5 for only a select list of undisclosed “trusted partner” organizations, before lifting the export controls entirely on June 30, 2026.

Then, on July 14, 2026, the White House, Treasury, the Department of Homeland Security, and the Pentagon jointly launched “GOLD EAGLE,” a standing industry-government coordination body with undisclosed power “to collectively identify risks, prioritize action, and strengthen the resilience” of AI systems, resting on no new statutory authority. Separately, OpenAI has agreed, through informal negotiation rather than any public order, to limit distribution of its own forthcoming models to government-vetted partners.

And this week, on August 1, 2026, the White House stated that it had finalized a “voluntary” framework for reviewing advanced AI models before their release, while declining to disclose the framework’s contents, which companies have agreed to participate, or when the process will take effect.

As the Justice Department’s former head of antitrust policy, David Lawrence, put it:

President Trump is planning product releases and choosing customers of a key industry in our ostensibly free-market economy.

Share

What the public needs to know — and why

The White House bypassing Congress to create its own secret AI regulatory regime could raise concerns in any context. It’s particularly worrisome in an era of authoritarian governance. There are three reasons why:

Autocratic capture. A common move from autocrats is to use the power of the state to capture key nodes of private industry to coerce them to do the bidding of those in power. Protect Democracy’s executive director and co-founder, Ian Bassin, termed this autocratic capture. It’s an inverse of regulatory capture where the regulated industry captures the state. We’ve seen extensive efforts from the Trump administration to capture the media so that it adheres to the administration’s party line.

The new regulatory framework gives the White House existential power over the AI companies — a critical component of the information landscape — and with that, immense leverage to engage in autocratic capture of this industry. We need to know what safeguards are in place to prevent the White House from using that sword-of-Damocles leverage in order to coerce the AI companies to do its bidding on any number of topics, from the use of deepfakes to how chatbots describe election results. As we’re already seeing, to implement the White House’s EO banning “woke” AI, the administration is ready to enforce its view of historical truth and neutrality as well as leverage federal procurement rules to decommission and punish contractors using models that display what the administration deems “partisan bias.”

Turning public office into self-enrichment. This administration has also gone to extraordinary lengths to convert its public office into personal self-enrichment for the president, his family, and those around him. Claiming unilateral control over one of the country’s largest and fastest-growing industries provides enormous opportunities for officials with conflicts of interest to act for their own financial benefit, not the public interest.

Already, there appears to be at least some degree of entanglement between the president and his family and parts of the AI industry. Recent reporting about Trump’s personal fundraising efforts raises serious questions about the financial relationships between AI companies and private entities in which the president has a financial stake, including Trump Media & Technology Group and MAGA Inc. As a Wall Street Journal analyst has described it, “we’re replacing the free market with the market for Trump’s affections.”

We need to know what safeguards are in place to avoid conflicts of interest in regulatory decisions — including what the applicable criteria are; who is applying them; and how we can know that they are operating without a personal stake in the companies they are regulating.

Coercive crony capitalism. Relatedly, this administration has demanded payoffs to support its pet and political causes — from the White House ballroom to the president’s political action committees. It is able to obtain these payouts because of its willingness to weaponize the power of government to reward friends and punish opponents. Reporting in the Wall Street Journal has indicated that President Trump has taken a close personal interest in how much companies are donating to his political interests, and “companies have paid new attention to the White House as Trump has gotten involved in regulatory decisions that once were made by independent agencies.”

That reporting includes AI companies and their customers.

The new framework gives the White House more leverage to demand money from AI companies, investors, executives, and employees. (And as we know, this is an industry that right now has a lot of money.)

Conversely, the secret executive regulatory regime also gives industry participants an opportunity to use undue influence to secure regulatory action that benefits them at the expense of their competitors — a classic crony capitalism dynamic.

A proper regulatory regime should have ample guardrails to mitigate the risk of regulatory power being leveraged for private political fundraising, or political contributions being deployed to pick winners and losers. The public needs to know if any such safeguards are in place here. Or, lacking those safeguards, in what sort of ways the president is seeking to solicit money from this industry — or parts of the industry are using undue influence and access for competitive advantage.

Whatever the public’s views on how artificial intelligence models should be regulated, we should all be concerned about a secret executive branch legal regime, especially from an administration that has exercised its power for political and personal gain. Much can go wrong when the executive branch bestows upon itself the power to decide, through undisclosed criteria and without congressional authorization, which companies succeed and which customers gain access to a general-purpose technology reshaping the economy. With an administration in office that wields its authority to entrench its power, enrich itself, and reward its friends, those concerns are greatly magnified.

A starting point to mitigating these concerns is to make public how this new legal regime works, who is involved in decisions, with what criteria, and how to avoid corruption and improper influence.

We’ll keep you posted on what our investigation turns up.

Next week marks one year since President Trump deployed National Guard troops to the streets of Washington, D.C. What began under the thin pretext of a “crime emergency” has devolved into a long-term — and unprecedented — military deployment in the nation’s capital.

In recent months, the deployment has ballooned to roughly 5,000 troops from D.C. and nearly two dozen states operating under Joint Task Force-District of Columbia. It’s been one dubious justification after another: the “crime emergency”; beautification of the Mall and other landmarks; security for high-profile events like the UFC fight at the White House. Abandoning the facade of exigent necessity, the Pentagon just confirmed that the D.C. mission will, at the very least, continue through January 2029 — when the next President is sworn into office.

Subscribe now

Roving patrols of armed military personnel have been a familiar sight in Moscow, Cairo, Havana, and other capital cities around the world going back decades. But in the United States, law and tradition have long excluded the armed forces from domestic policing, ensuring the military could focus on its actual mission: national security.

Yet armed Guardsmen are now a routine sight throughout the District, rarely prompting headlines or national outcry — apart from exasperated D.C. residents who condemn the “occupation” of their neighborhoods. As fresh outrages have taken over the headlines, the prolonged deployment normalizes the day-to-day presence of uniformed troops policing civilian life. Compounding this problem was the decision by several Democratic governors to — after previously arguing in court that the mission was dangerous and illegal — accept the administration’s questionable public safety claims and deploy hundreds of their Guardsmen to D.C. during this summer’s Freedom250 programming.

The ripple effects of an indefinite, overtly political deployment extend far outside of the District, undermining civil-military relations, weakening the military’s nonpartisan ethos, and inviting future authoritarian abuses. Notwithstanding the integrity and discipline of our men and women in uniform, these risks are amplified by Defense Secretary Pete Hegseth’s comprehensive efforts to politicize the armed forces — from staging tank parades for the president’s birthday, to staging political rallies on military bases, to forcing senior officers into early retirement for exercising independent judgment, to corrupting the promotions process with a discriminatory political ideology.

These emerging dangers explain why separating military operations from domestic law enforcement has been a core tenet of the U.S. constitutional order, pre-dating the Revolution with hard-earned lessons from life under British military rule. Two hundred fifty years of self-governance have reaffirmed the wisdom of this principle. So have countless other episodes around the world.

We recently published a report — The Dangers of Military Intervention in Civilian Law Enforcement — on military interventions in domestic policing across history. We found that, time and time again, they produce three dynamics that are dangerous to civil society, the rule of law, and the military itself: escalation, disillusionment, and politicization.

One year in, and all three trends are clearly visible in D.C.

Read our report

The deployment hasn’t reduced violent crime, but it has raised the risk of escalation

One recent analysis by the Center for American Progress finds that the National Guard’s D.C. deployment has had no measurable impact on the reduction in violent crime that started in earnest back in 2024. The Niskanen Center reached a similar conclusion, characterizing the deployment as “an expensive tool deployed in the wrong places for the wrong types of crime” with “no measurable effect” on violent crime in the District.

The fact that, for much of the past year, Guard patrols have predominately targeted the Mall and other high-traffic tourist areas rather than neighborhoods facing the highest crime rates underscores the White House’s prioritization of optics over genuine efforts to improve public safety.

Meanwhile, the escalatory risks to troops and civilians alike have grown. In November 2025, a targeted attack on two West Virginia National Guard members tragically killed Specialist Sarah Beckstrom and critically injured Staff Sergeant Andrew Wolfe. At the start of the deployment, military commanders had already contemplated the heightened risk of placing Guard members in harm’s way with their highly-visible D.C. patrols. After the attack, one D.C. Guardsman reflected: “Those poor kids just came to the city, and they didn’t want to be here … The whole thing is … senseless … I gave up trying to understand it.”

Tensions between D.C. residents opposed to the deployment and the Guard have continued to escalate in recent months. Though the Guard lacks law enforcement authority, the U.S. Marshals Service has deputized troops as special police, allowing them to carry firearms and detain individuals until police officers make any formal arrest.

In May, Guard personnel reportedly held down and detained Amy King, an Army combat veteran, outside of her D.C. home. When D.C. police arrived at the scene, an officer reported that King was involved in a prior confrontation — an allegation that King categorically denies. Though the full picture isn’t entirely clear yet, King filed a legal claim to challenge the Guardsmen’s apparently excessive force. This episode again emphasizes the fundamental mismatch between the Guard’s training and the de-escalation skills required during civilian confrontations — a gap that makes even routine encounters between troops and civilians more likely to escalate.

The costly deployment has a mandate far beyond the Guard’s core mission, breeding disillusionment

Polling consistently shows that the D.C. deployment is unpopular among residents (and Americans nationwide), fueling disillusionment inside and outside the ranks as troops take on duties well beyond the scope of their traditional mission.

The deployment’s substantial focus on groundskeeping, sanitation, and other “beautification” efforts around the District bears little resemblance to the troops’ specialized training and operational expertise. Reflecting that mission creep, a congressional report notes that the D.C. Guard has more specified metrics for measuring its beautification efforts than for measuring the deployment’s impact on public safety. Even Acting Attorney General Todd Blanche framed the mission through these efforts, telling a crowd of Guard personnel they should measure their success “[N]ot just [in] the reduction in crime, but the beautification of the city … [and] the cleanliness of the city.”

And the price tag only keeps growing. The deployment is already on track to cost more than the Metropolitan Police Department’s (MPD) entire annual operating budget, and could total between $2.5 billion and $3.4 billion by January 2029. The Pentagon’s own more conservative estimate, which assumes a troop drawdown from current levels, still puts the total well over $1 billion. The administration seems unfazed by the mounting costs, and officials recently signed a $292 million contract to house Guard personnel in what the National Guard Bureau called “the largest interagency lodging acquisition” in the Guard’s history — for a deployment that most local residents neither asked for nor want to continue.

The costs of the prolonged deployment extend to the Guard’s readiness. D.C. Guard leadership reportedly expressed “significant concerns” that the deployment has already degraded — and will continue to degrade — operational readiness, particularly “in the event of a foreign conflict,” hardly a far-fetched risk as hostilities with Iran continue. The deployment has also strained the MPD’s own force readiness, pulling dozens of police officers off their beats to deploy alongside the D.C. Guard in lieu of their full-time law enforcement duties.

The administration has politicized the deployment

In the early weeks of the deployment last year, internal National Guard documents warned of growing perceptions that the mission would drive a “wedge between citizens and the military,” pointing out the predictable byproduct of a deployment tied to the administration’s power grabs. Whatever the Guard’s efforts to remain above the fray, domestic deployments driven by politics rather than genuine security needs often perpetuate a cycle of politicization. As the D.C. deployment has become more embedded over the past year, reports of Guard personnel veering into politically fraught territory have also increased.

Ahead of the D.C. primary elections in June, footage showed Guard personnel patrolling around a ballot drop box, immediately raising concerns about potential voter intimidation. The fact that troops apparently received no clear instructions before voting started to avoid areas around polling sites or drop boxes shows how prolonged deployments have the potential to create downstream political risks.

Another video that was posted to the Guard’s website — before it was removed following public backlash — showed departing Arkansas Guard members signing a “Free D.C.” flag as if it were a trophy or souvenir from battle. Whatever the intent, the optics alone show how the continued deployment of out-of-state troops fuels further politicization that ultimately undermines the Guard’s apolitical tradition.

Earlier this summer, Guard personnel were roped into President Trump’s Reflecting Pool saga, patrolling the grounds around the Lincoln Memorial alongside U.S. Park Police in search of alleged vandals and threatening to detain anyone suspected of touching the water. And in July, hundreds of Guard members served as the staged audience for a political stunt by Secretary Hegseth and Stephen Miller, placing troops squarely in the middle of the administration’s politicized rhetoric aimed at D.C. residents and those opposed to the deployment.

Taken together, these developments demonstrate how the deployment’s continuation normalizes the Guard’s presence on America’s streets — accelerating the potential for more overt politicization down the line — in D.C. and elsewhere. Normalization invites mission creep, and an ever-expanding mandate will inevitably serve political ends. This is already happening: the Tennessee Guard has been deployed for months in Memphis as part of a federal task force focused principally on advancing the administration’s domestic immigration agenda.

When the military is used for political purposes, public trust, operational readiness, and civil-military boundaries all erode. And a military tapped to carry out a partisan agenda on behalf of whoever holds power threatens the very constitutional order it is meant to protect.

There are still opportunities to change course

With the administration’s continued extension and expansion of the National Guard’s deployment in Washington, D.C., it’s increasingly clear that normalizing the presence of armed troops patrolling the nation’s capital is a feature, not a bug, of the mission. But we need not acquiesce and march down this troubling path. There are steps we can all take to ensure that the past year in D.C. does not become the new normal.

• Tell your governor to keep the Guard at home. Roughly 90% of the Guard personnel currently in D.C. come from states around the country — Alabama, Alaska, Arkansas, Florida, Georgia, Hawaii, Idaho, Iowa, Louisiana, Michigan, Mississippi, Nevada, South Carolina, West Virginia, and Wyoming — and governors can freely decline participation or recall deployed troops at their discretion. Indeed, roughly half of sitting governors — of both parties — have refused. Others have brought troops home early after learning more about the situation on the ground. Deployments to D.C. take employees out of their full-time jobs, parents away from their families, and undermine the mission readiness of Guard units for core functions at home and abroad. The risk of serious political backlash can help governors muster the necessary courage to resist or draw down existing deployments.

• Tell your Members of Congress to demand accountability. One of the few must-pass bills every year authorizes our national defense policy, providing lawmakers the opportunity to rein in the administration with provisions limiting domestic military deployments and demanding transparency from administration officials. Provisions tackling other areas of executive branch overreach have already drawn bipartisan support this cycle, and the legislative door for reform remains open not just this year but in the years ahead.

• Distinguish between honorable troops and dishonorable policy. The overwhelming majority of Guardsmen deployed to D.C. have followed the law and treated civilians with respect. Many themselves have misgivings about the mission and are doing their best to navigate difficult and unpleasant circumstances. The public absolutely should demand accountability for individual incidents involving excessive force or other unlawful conduct. But advocacy and rhetoric should be careful to avoid alienating law-abiding troops and their families, when it is civilian leaders who bear ultimate responsibility for this unnecessary and improper deployment.

• Veterans and retired national security leaders can speak out. These voices carry unique credibility on the threat that improper domestic deployments pose to healthy civil-military relations, and many have used their platforms to inform and persuade. Many joined a legal brief explaining the dangers of the D.C. mission to the health and strength of the military. Others have spoken up in support of the governors who declined to send troops this summer.

This piece was originally published by The UnPopulist.

At LibCon2026’s “Checking the Imperial Presidency” panel two weeks ago, the panelists reacted vociferously to a question about expanding the Supreme Court. Supporters of court expansion see it as a suitable response to the court’s anti-democratic rulings; the panelists unanimously insisted that expansion would put the nail in our democracy’s coffin by triggering a tit-for-tat downward spiral. (The UnPopulist’s Andy Craig has offered a proposal for a court expansion that would repudiate the current court’s majority while offering an off-ramp of a constitutional amendment.)

And just like that, what could have been a serious conversation about the Supreme Court’s role in democratic backsliding and how to respond got buried in an internecine dispute.

Subscribe now

That’s not to blame the panelists: If anyone’s to blame, it’s the moderator. Which is to say, we should look in the mirror. One of us (Justin) moderated that panel; the other (Erica) helped prep for it. And in hindsight, here’s what we wish we could have injected into the conversation:

This panel’s name — Checking the Imperial Presidency — elides an important partner in Trump’s project to transform our democracy into a more authoritarian form of government: the Supreme Court, which under John Roberts has acted as Trump’s consigliere. Through parallel abuses of power, Trump and the court have launched a two-branch doom loop, with mutually reinforcing moves to enhance their own power, render themselves unaccountable, entrench an imperial presidency and uncheckable court, and sideline the people’s representatives in Congress.

We cannot rebuild our democracy and the rule-of-law — including constraining the imperial presidency — without solving the problem of the Roberts Court.

But how to do that?

As at the panel, the conversation often jumps straight to court expansion, whose perils and promises have been extensively debated. So too with term limits, which are a somewhat less controversial idea.

But those reforms need not be the only options on the table.

There is a range of reforms that Congress can enact to take back power from the Supreme Court (and the presidency). And indeed, the problem of the Supreme Court in our democracy extends beyond which justices control the court. The system of checks and balances does not exempt the Supreme Court, and Congress has the tools to check the court just as the court has tools to check Congress.

But first, let’s examine the stakes: why it’s not enough to talk about the imperial presidency anymore. We need to talk about the two-branch doom loop and the mutually reinforcing, self-aggrandizing moves from Trump and the court.

The Trump presidency and the Roberts Court: A vicious cycle for democratic backsliding

Trump’s arsenal of power-grabbing moves is well known, the Supreme Court’s less so. Often those who claim that the court’s record is not facilitating democratic backsliding point to its pushback on particularly egregious — and obvious — Trump excesses such as abusing emergency powers to impose tariffs. But it is worth taking a more comprehensive look at its record, including at some of its lesser-known rulings, to get a fuller picture. The court does not have to be the president’s “puppet” for their commitments to align and facilitate a destructive cycle.

1. Like Trump, the court has operated in secrecy and stonewalled oversight, treating transparency to Congress and the American people as a threat to be neutralized. The court has shielded itself from transparency, through refusing to explain the basic reasoning of its momentous rulings on the emergency shadow docket. It has stonewalled congressional oversight on matters ranging from cybersecurity practices to corruption among the justices. And it has reinforced the secrecy of the Trump administration, making it harder for Congress to investigate Trump and protecting him from the oversight associated with criminal prosecution.

2. Like Trump, the court has declared that laws preventing corruption don’t apply to it. Supreme Court justices have flouted the ethical constraints that bind every other federal judge, accepting millions in undisclosed gifts and travel from parties with interests before the court, refusing to recuse in the face of clear conflicts of interest, and resisting for years the adoption of any binding ethics code. In so doing, they appear to have violated federal laws — all while suggesting that Congress cannot regulate to prevent their corruption. And the court’s decisions gutting corruption laws, creating presidential immunity, and curtailing congressional oversight of the White House have enabled the Trump administration’s corruption.

3. Like Trump, the court has exploited emergency authorities to expand its power. The Roberts Court has weaponized its emergency or “shadow docket,” granting last-minute stays that effectively decide major policy questions about how our government works and the scope of people’s rights without full briefing, argument, or reasoned opinion. This has made it difficult for lower courts to even know what rule they are supposed to follow when a case is sent back down, to the open frustration of judges. The Roberts Court’s decision not to pause the Trump “emergency” tariffs demonstrates its use of its emergency authorities to reinforce the president’s emergency authorities, even when it ultimately holds the presidential acts unconstitutional.

4. Like Trump, the Supreme Court has consolidated complete power over its own branch. The Roberts Court has reserved to itself the power to issue universal injunctions, while — at the behest of the Trump administration — moving aggressively to curtail the power of federal district courts to do so. That, combined with its enthusiasm for the shadow docket, further consolidates power in this single court, rather than the federal judiciary as a whole. And the court has blessed the Trump administration’s strong unitary executive view that the president is the executive branch (l’état c’est moi) in order to claim control over the DOJ, gut independent agencies, fire inspectors general, and purge the civil service.

5. Like Trump, the court has manufactured false “facts” to enable it to achieve its desired outcomes. The court has consistently made up “facts” of its own. Even those less perturbed by the outcome of some rulings should be troubled by the court’s methods, which violate the principle that the court is supposed to be deciding legal questions, not doing its own fact-finding. In Dobbs, it “misrepresent[ed]” history. In Shelby County, it overrode the findings of more than 20 congressional hearings and 15,000 pages of documents with its own vibes-based assessment of racial discrimination in covered jurisdictions. And in Kennedy v. Bremerton School District (the football team prayer case), the court hallucinated its own record. The Roberts Court, at the same time, has often ratified the Trump administration’s objectively false claims.

6. Like Trump, the court has manipulated the rules of democratic participation to entrench partisan power and avoid democratic accountability. The Roberts Court has consistently acted to entrench the power of the Republican Party. It has, for example, gamed the “Purcell principle” for maximum partisan advantage (this is the general rule that federal courts should refrain from altering election rules on the eve of an election to avoid confusing voters and burdening administrators); gutted the Voting Rights Act to the advantage of its partisan allies; and protected the Republican presidential candidate from laws enacted by Congress. In the presidential immunity case, the court deployed the rhetoric of presidential maximalism while denying the sitting Democratic administration the very authorities, conclusive and preclusive power to make prosecutorial decisions, that it insisted President Trump had every right to claim.

7. Like Trump, the court has dispensed with (ignored) laws it dislikes, rather than faithfully executing or applying them. Under the cover of “constitutional avoidance,” the Roberts Court has repeatedly rewritten statutes to achieve its preferred policy outcomes. Its approach to statutory interpretation appears inconsistent and hypocritical. The court’s amorphous “we know it when we see it” standard resembles its own policymaking more than a court announcing legal rules. For example, the court held that a decision by the Department of Education to cancel student loans presented a “major question” that only Congress could weigh in on but that dismantling the entire department did not.

8. Like Trump, the court has arrogated to itself the power to set the national policy agenda. The Roberts Court has used its agenda-setting power to facilitate many of Trump’s executive power grabs and separately to upset congressional compromises on health care, consumer protection, agency structure, voting rights, civil service protection, and campaign finance.

9. Like Trump, the court has sought to have the last word, so no democratic response can dislodge them. The court has structured its most consequential decisions to foreclose response: striking down campaign finance laws, statutes structuring government agencies, and voting rights protections on grounds that make legislative correction practically impossible. For example, its decision striking down part of the Voting Rights Act in Shelby County was followed immediately by a wave of restrictive state voting laws, entrenching political power before Congress could respond. Cases that are nominally about statutory interpretation and thus deference to Congress often operate in the political reality that congressional response is unlikely and can take many years if it comes at all.

The Trump administration and the Supreme Court have thus each used the other to ratify, reinforce, and accelerate its assaults on our republican form of government and Congress’s authority.

In one variation of this cycle, Trump has been the first mover, embracing a radical approach to his own power; the court has then validated, extended, and constitutionalized these moves, and Trump has taken the new runway and pushed even further. For example, in Trump v. United States, Trump made an unprecedented claim to absolute immunity from federal criminal prosecution for many official actions; the court granted the immunity and called it constitutional law, in part basing this immunity on presidential control over the DOJ and thus the president’s need to be free from any interference in his commands to subordinates. Then, the second Trump administration took advantage of the court’s blessing to fully weaponize the DOJ and engage in extreme abuses of power.

Other times, the court moves first, imposing its own partisan-entrenching policy preferences. Trump then pushes to further Republicans’ partisan advantage, and the court ratifies the move. The attack on voting protections has followed this model: Shelby County and Rucho closed off protections for voting rights and against gerrymandering. Trump then pushed for mid-cycle extreme gerrymandering and for gutting the rest of the VRA, and the court granted this wish and wrote it into law.

Taken together, these power-aggrandizing moves from Trump and the Roberts Court create a doom loop that is on the precipice of enshrining authoritarianism into our Constitution.

Congress can check the Court just like it can check the presidency

The conventional wisdom about Supreme Court reform is stuck in a rut. Most reform advocates focus either on ethics or on changing the court’s composition through term limits or expansion. But the problem is not just who sits on the court, or whose money they take in their extracurricular activities. It is how much power the court exercises, with how little accountability, over how wide a range of questions. Without structural constraints on what any court can do, any reform will just be a band-aid.

What follows is a different approach: a menu of possible reforms aimed at disempowering the Supreme Court as an institution — constraining how it acts, requiring it to show its work, forcing it to share power with the rest of the federal judiciary, and giving Congress meaningful tools to respond. None of these reforms is exotic or unprecedented. (Professors Daphna Renan and Niko Bowie’s forthcoming book gives numerous examples of when Congress has turned to these types of solutions in American history.) Most have analogues in the tools Congress should use to check the executive branch. Readers may find some more compelling than others. They do not all need to be adopted. But some combination of them would restore the court to its proper role as one participant in constitutional governance rather than its unelected sovereign. Enacting these reforms without more will not be the end of the process. The Supreme Court will surely try to fight back by ignoring or invalidating them, at least to some extent. And Congress will need to use the full scope of its authorities in response.

But they’re a necessary starting point to escape the two-branch doom loop.

Reform the shadow docket. The Roberts Court has remade vast areas of American law — abortion access, immigration enforcement, environmental regulation, election rules — not through the deliberate process of full briefing, argument, and reasoned opinion, but through unsigned emergency orders issued in the middle of the night with no explanation and no accountability. These orders are analogous to executive rulemaking, but unlike agency rules they require no notice, no comment, no reasoning, and no disclosure of who voted for what. Congress should require written opinions for any stay of lower court proceedings, signed by at least five justices; impose mandatory waiting periods before emergency orders take effect other than in true emergency situations (such as an immediately pending execution of capital punishment); and mandate public disclosure of how each justice voted. This is not a radical constraint on judicial power—it is the same transparency and procedural discipline Congress has long required of executive agencies under the Administrative Procedure Act. If agencies must explain their emergency actions, so must the court. Alexander Hamilton’s famous observation that the judiciary is the “least dangerous” branch was premised on a court’s need to explain its reasoning, which goes out the window when the court simply refuses to do so.

Require evidence-based adversarial fact-finding. One of the Roberts Court’s most insidious practices is making up facts. In case after case, these are not legal conclusions — they are factual claims about the world, and they are being made by the institution with the least procedural accountability of any factfinder in the legal system. The court does this even though, outside of rare original jurisdiction cases, the court’s appellate function is supposed to be deciding questions of law. Developing the factual record is supposed to happen at the trial court level, and the Supreme Court’s contortions to get around this are increasingly strained. Congress should require that any factual proposition — case specific, social-scientific, economic, or historical — that is material to striking down a statute be developed through adversarial proceedings at the district court level, with expert witnesses subject to scrutiny and cross-examination, and reviewed by the Supreme Court only for the traditional standard of a facially obvious “clear error.” The same Congress that requires agencies to base major rules on substantial evidence developed through open adversarial processes can require the same of the court.

Require the court to show its legal work. When a federal agency acts, it must identify its legal authority, explain why it chose the action it took, and address contrary evidence in the record — or a court will reverse it. This is the Chenery principle, one of administrative law’s foundational constraints on executive power, and the Roberts Court has applied it enthusiastically when it suits its desired outcomes. It does not apply this to itself. The court routinely invalidates or rewrites statutes without explaining why no narrower reading was available, which constitutional provision it is applying, or how its interpretation squares with the historical evidence it claims to rely on. Congress should require, as a condition of invalidating federal legislation, that the court’s opinion do each of these things. This reform does not tell the court how to rule — it tells the court to explain its rulings with the same rigor it demands of every other government actor.

Hold justices to the professional employment standards that govern modern workplaces. In addition to deciding cases, Supreme Court justices are bosses. But unlike most other attorneys and government officials, they supervise employees free of the constraints of employment laws, and they run hiring processes that are notorious for identifying politically aligned and politically engaged law clerks who serve for one year and then are dependent on “their” justice as they embark on their careers. Whatever arguments support ideologically aligned hiring for some jobs in Congress and the executive branch wilt at the courthouse doors. If justices are to be something other than politicians in robes, the workplaces they run need to reflect that. Options for regularizing the workplace include making Supreme Court clerks the equivalent of short-term Schedule A executive branch employees (the line-level attorneys at the Solicitor General’s Office, Civil Appellate, and the Office of Legal Counsel are typically Schedule A), with associated rights against workplace discrimination; creating a pool of clerks that rotate through the different justices’ chambers (analogous in some ways to the Supreme Court’s cert pool); or reducing the number of clerks that each justice can hire while bulking up the staff attorney pool at the court. Congress could also potentially require the hiring clerks to be conducted more at the court-wide level, under a standardized process, and less by individual justices.

End the corruption and make ethics rules enforceable. Supreme Court justices have accepted millions in undisclosed gifts and travel from parties with interests before the court, refused to recuse in the face of clear conflicts of interest, and for years resisted any binding ethics code — all while justices have suggested that Congress cannot regulate to prevent their corruption. Every other federal judge is subject to the Code of Conduct for United States Judges and enforceable recusal standards. The Supreme Court’s exemption from those norms is not required by the Constitution or by any serious theory of judicial independence. It is an anomaly the justices created for themselves. Congress should establish mandatory recusal standards enforceable by a panel of senior circuit judges, require financial disclosure at parity with lower federal court judges, and create a formal mechanism for parties to seek recusal with written rulings. The same Ethics in Government Act logic that applies to executive branch officials applies here. The court is not above the law it administers. Even if justices aren’t literally selling their votes in individual cases, the purpose of recusal and binding ethics codes is to protect public confidence by avoiding improper influence or even the appearance of conflicts of interest. Flouting these standards has seriously undermined the court’s credibility.

Require broader consensus before the court can erase an Act of Congress. Right now, a bare five-justice majority can permanently nullify legislation enacted by overwhelming democratic majorities — the Voting Rights Act, the Affordable Care Act, campaign finance law. It can also effectively amend the Constitution by declaring new constitutional rules that it then applies to laws Congress has previously enacted. The Constitution already recognizes that some actions are too consequential for bare majorities: two-thirds of the Senate must ratify a treaty, two-thirds of both chambers must override a veto, two-thirds must convict an impeached official. And amending the Constitution requires ratification by three-quarters of the states. The same logic applies to judicial amendment of the Constitution or nullification of democratically enacted law. Congress should require a super-majority of six or seven justices to agree before it invalidates a federal statute.

Channel cases to appropriate courts. Congress could direct that certain categories of high-stakes constitutional litigation — challenges to the structure, spending, and staffing of the federal government; suits seeking to invalidate an Act of Congress; and disputes over voting and elections — be heard in the first instance by specialized three-judge panels rather than single district judges. This is not a novel mechanism: Congress has long required three-judge district courts for redistricting and other voting-rights cases, most recently in the Electoral Count Reform Act of 2022, precisely because of their importance and time sensitivity. That existing framework could be extended to cases challenging the basic architecture of federal governance. As for the path beyond the three-judge panel, Congress retains the choice: it could preserve the Supreme Court’s ultimate appellate jurisdiction over these channeled cases, or it could instead make the panel’s judgment final for a defined category of claims, exercising its Article III, Section 2 authority to except certain classes of cases from the court’s appellate jurisdiction altogether. This is the “jurisdiction-stripping” approach'; Congress has done it before, and the court has affirmed this congressional power. Another approach would be to set a narrow time window within which courts can review the constitutionality of legislation, so that if the court strikes it down Congress can quickly respond.

Give Congress a meaningful voice in constitutional interpretation. The Roberts Court has structured its most consequential decisions to foreclose democratic response — striking down voting rights protections, campaign finance laws, and agency structures on grounds that make legislative correction practically impossible, then watching as the political beneficiaries of those rulings consolidated power before Congress could act. A partial solution is fast-track voting procedures for Congress to respond to statutory decisions by the court and clarify what it meant the statute to do. But even on constitutional decisions to invalidate statutes, Congress could have a real opportunity. One approach would provide that if consecutive Congresses re-enact a statute the court has struck down, with an intervening election, that re-enactment is protected from judicial review. That would ensure there is a sustained majority backing Congress’s constitutional view, and limit situations where Congress has the last word in a moment of political passions. (This could be modeled after the law for, e.g., amending the Virginia Constitution.) Another variation, modeled on the Canadian “Notwithstanding Clause,” would allow a super-majority of Congress to re-enact a statute the court has struck down. Judicial review would remain a genuine check — the court’s ruling would stand unless Congress mounted a demanding supermajority response. In either variation, constitutional meaning would become a dialogue rather than a judicial monologue, and Congress would recover the ability to respond to constitutional decisions about the laws it enacts.

None of these reforms eliminates judicial review entirely. None eliminates judicial independence. None engages in partisan power plays. None threatens the ability of individuals to vindicate their rights in court. None requires a constitutional amendment. What they share is a recognition that the Roberts Court’s power is not self-executing or constitutionally required at its current scale — it is a product of choices the Roberts Court itself has made and that the Constitution leaves to Congress to revise. The solution to a court that has accumulated too much power is the same as the solution to a presidency that does the same: Congress, doing its constitutional job.

This piece was originally published by The UnPopulist.

In the 2024 elections, only about four percent of U.S. House seats changed hands from one party to another. Four percent. It doesn’t seem possible to make our political system even more rigid. But the two major parties are trying anyway.

In advance of the 2026 midterms, state legislators are racing to redraw congressional maps to make seats safer for one party or the other. To make matters worse, the Supreme Court’s recent decision in Louisiana v. Callais effectively gutted Section 2 of the Voting Rights Act — the provision that for decades allowed voters of color to challenge maps drawn to dilute their voting power.

Subscribe now

Most Americans live in districts where the outcome was decided long before Election Day by politicians drawing lines to protect their seats. This November, millions of voters will cast ballots that are powerless to affect those outcomes. With this latest wave of mid-cycle redistricting, we’re entering a new phase of a problem that’s plagued American democracy from nearly the beginning.

It’s also a problem that most other democracies have managed to solve.

Gerrymandering persists because single-member districts allow it

In 1968, the political scientist Robert Dixon quipped that “all districting is gerrymandering.” Dixon was making an observation about a fundamental feature — and flaw — of the U.S. electoral system itself: our use of single-member districts.

Single-member districts are uniquely vulnerable to manipulation. Gerrymandering works by exploiting a common feature of them known as “wasted votes.” When a candidate wins with 60% of the vote, the 40% cast for other candidates doesn’t help to elect anyone; those votes are considered “wasted.” So, too, are the 10% the prevailing candidate earned beyond what was needed to win. Gerrymanderers are essentially engineers of this waste, designing districts to maximize the opposition’s wasted votes and minimize their own.

Read more — How to end gerrymandering

Take Austin, Texas, as an example. In 2011, mapmakers sliced up the city by drawing parts of it into several surrounding districts. This dispersed the city’s largely Democratic voters across multiple districts. Republicans were able to win in race after race by small but comfortable margins (minimizing waste) while Democrats were spread thinly across the map (maximizing waste). The result: the next year, Republicans in Texas won roughly 58% of the statewide vote but secured 67% of the state’s House seats. Of course, this cuts both ways. In 2024, Democrats in heavily gerrymandered Illinois won about 53% of the vote but ended up with 82% of the state’s seats.

That each of our congressional districts elects only one representative seems natural enough — by and large, that’s the way it’s always been. But globally, this is unusual. The reason we so rarely hear about gerrymandering elsewhere is that most democracies use a system where gerrymandering simply doesn’t work.

Proportional representation inhibits gerrymandering at the source

Today, Protect Democracy released a new report, “When Maps Don’t Matter,” that unpacks how adopting a single reform for the U.S. House would end the redistricting wars for good.

Read more — When Maps Don’t Matter

Most democracies now use multi-member districts where winners are allocated proportionally to votes: a system we’ve long supported called proportional representation. In a five-seat district, a party that wins 40% of the vote receives two seats, and another that wins 60% receives three. As districts grow in size — from one to three members, or from three to five — the ability to manipulate outcomes through line-drawing rapidly diminishes.

Proportional systems aim to align seats with votes: winning about half the vote gets you about half the seats — not 82%. Try as they might, would-be gerrymanderers can’t do much about it, because proportional systems deprive them of their one key tool: wasted votes. Recall that a party that wins 40% of the vote in a single-member district wins nothing (lots of waste). But in a five-seat district, that same party wins two seats: instead of being wasted, nearly all of the votes contribute to representation.

Consider what happens when mapmakers try to gerrymander under proportional representation. Certainly, one could move voters across a district line. But doing so just means those voters bring their representation with them. If 40% of voters are shut out of representation in one district, they simply elect their proportionate share somewhere else. Shuffling the map just shuffles where their seats come from, but ultimately leaves them with the same amount of representation.

Sign up to learn more about proportional representation

Proportional systems make gerrymandering prohibitively difficult for another reason: It tends to generate multiparty competition. In a two-party system, mapmakers can often make educated guesses about how voters will split between the two sides and draw districts accordingly. But when three or five or more parties are competing, those calculations become much less predictable. A smaller party might surge unexpectedly, drawing support from several different parties and reshuffling the political landscape. Multipartyism increases uncertainty — and uncertainty is the enemy of gerrymandering. When it’s harder to predict outcomes, it’s harder to manipulate the maps in advance.

Over 500 scholars call on Congress to end the redistricting wars through proportional representation

Ridding our politics of gerrymandering will ultimately require reforming the system that enables it.

So say (and we agree) more than 500 scholars — including the nation’s leading democracy, voting rights, and redistricting experts — who also today released an open letter calling on Congress to adopt proportional representation as an end to the redistricting wars.

Read the full letter

“These developments are serious,” wrote the scholars, referring to recent developments like the Supreme Court’s decision in Callais and the escalation of mid-cycle redistricting. “But they should not obscure the underlying diagnosis: gerrymandering persists because single-member districts allow it. We write to urge Congress to do what most democracies have already done: adopt proportional representation.”

Reforms that preserve single-member districts are not enough

It’s possible (sometimes) to get rid of intentional gerrymandering while leaving single-member districts intact. But here’s the catch: Lopsided results tend to persist anyway.

In the 2024 election, California used maps drawn by an independent redistricting commission. Under these maps, Democrats won roughly 60% of the statewide vote but still earned about 83% of the state’s U.S. House seats. This gold-standard commission had all but eliminated gerrymandering, but Democrats still won far more seats than their vote share would suggest. What happened?

In California, Republicans tend to be dispersed geographically, spread out across regions dominated by Democrats. Despite the commission’s best efforts, this prevented mapmakers from drawing districts that would group Republicans together and allow them to win seats commensurate with their actual numbers. This problem is known as geographic sorting. Even in states where gerrymandering isn’t much of an issue, single-member districts are still uniquely sensitive to voter geography — producing distorted results similar to those generated by gerrymandering.

To be sure, reforms like independent commissions or proposed bans on gerrymandering are an improvement over the status quo. California’s results would almost certainly be even more lopsided otherwise (as they will likely be again, now that the state has voted to pause the commission). But as the more than 500 scholars explain, the use of single-member districts “makes the drawing of district lines consequential in the first place. As long as that system remains intact, distorted election outcomes will persist.”

It’s time to eliminate the problem at its source

Americans have never been fond of gerrymandering.

The term itself was coined to mock a salamander-shaped district drawn to protect the party of Massachusetts Governor Elbridge Gerry in 1812. Over 200 years later, Americans continue to deride the practice. In a recent Economist/YouGov poll, roughly seven in 10 said states should not be allowed to draw districts to favor one party.

And over 200 years later, the problem is still with us — for the simple reason that our system makes it possible. It’s time for a new system that eliminates the problem at its source.

Sign up to learn more

In the last few months, the Department of Justice (DOJ) charged protesters in Minnesota with engaging in a criminal conspiracy, secured sentences of a combined 450 years in prison against protesters in Texas, reindicted former FBI Director James Comey over an Instagram post, and opened investigations into nonprofits it claims fund political violence. While each case is framed as a response to the threat of political violence, some seem to have little to do with it at all.

Subscribe now

These cases share a pattern: the Trump administration is using allegations of political violence as cover to violate individual rights, chill dissent, and intimidate civil society. Direct connections to actual acts of political violence seem to be, at best, a secondary concern. Taken together, this is a classic authoritarian playbook tactic — a blatant attempt to tilt the playing field in their favor ahead of the midterms, borrowed from the same script used to label nonprofits as foreign threats abroad.

The strategy could work because Americans are genuinely worried. Political violence has been on the rise for a decade, and now includes violence that comes from the government itself. Americans know that things are trending in the wrong direction, but polling shows they lack the full picture and they don’t agree on where the threat is coming from. That gap between fear and understanding is exactly what the federal government is exploiting.

Read more — Media Guides: Covering political violence

The data doesn’t match the narrative

The Trump administration has baselessly claimed the country faces “an organized campaign of radical left terrorism” that has broad support. That characterization may fit their narrative, but it doesn’t fit the data.

Experts broadly agree that threats and violence have increased in recent years. While high-profile attacks have come from both the left and right, most documented episodes over the last decade have come from individuals seemingly identifying with the political right. The number seen as coming from the left has begun to rise only recently, from a prior baseline of almost zero. And in fact, ideology itself is a poor lens for this problem altogether: Many perpetrators of political violence hold mixed or incoherent worldviews that don’t really map onto our conventional notions of left and right at all. The administration’s claims of organized, ideological violence campaigns just don’t reflect the realities of a serious and complex problem.

What is clear is that Americans across the ideological spectrum reject political violence. Just 16% of Americans think violence for political ends is ever acceptable, and only 2% are confident that they themselves would personally consider participating. These numbers are consistently low across the partisan spectrum — the differences between self-identified Republicans or Democrats are minor because support is so low across the board.

Covering political violence requires principle and clarity

Careful journalism is how most Americans make sense of the gap between perception and reality — but even good faith reporting can unintentionally stoke fear, push targeted communities from public life, or feed authoritarian narratives. This is why Protect Democracy and Over Zero have developed a series of resources to help newsrooms cover instances of political violence responsibly. How do we know if our news sources are living up to best practices? Three questions are worth asking of any story:

1. Is the reporting concrete, or does it sensationalize? Responsible reporting gets the facts right while avoiding language that stokes fear, misrepresents scale, or obscures accountability. If coverage heavily features natural disaster or war metaphors, or leaves you feeling powerless in the face of events that feel uncontrollable and unattributable, those are red flags. Precise numbers beat vague language. “Three protestors threw rocks at police” conveys a precise scale; “violence exploded” does not. It also avoids mistakenly implying that responsibility or support for violence extends to an entire group.

2. Does the reporting provide context? Reporting should note whether a targeted community has a history of being scapegoated and whether the event is connected to broader political narratives. Relying on experts for context is also important. Extremists use violence as a tool to amplify their reach, so reporting should never provide an unfiltered platform for violent actors and their ideologies. When a violent event involves agents of the state, journalists cannot limit their coverage to the immediate episode. They must explore whether and how the incident was part of a larger set of operations or events, and whether anyone faces accountability. Context is also critical for reporting on government responses. Violence can serve as a pretext for infringing on rights and freedoms or intimidating political opposition. Reporting should indicate when violence is being used to justify authoritarian crackdowns in the name of restoring “law and order.”

3. Does the reporting include the people targeted, their response, and additional condemnations of the violence? Political violence frequently aims to push targeted groups from public life — from voting and running for office to attending everyday community events — and to make everyday community members feel powerless in the face of repression. It has historically targeted Black, immigrant, LGBTQ, Jewish, Arab, and Muslim communities. But the threats to these communities are not the entire story. Communities take concrete steps to repair harm in the face of violence and protect public participation, particularly for those targeted. Reporting should cover these as well, taking particular care to engage with and amplify the perspectives of those targeted. Coverage should also highlight condemnations of political violence, especially from voices whose politics might otherwise seem to align with those responsible. Condemnation from an unexpected place reaffirms how widely violent tactics are rejected — and that itself is important context.

The stakes rise with the midterms on the horizon

Violence is a tool for disrupting elections. It can create an environment of fear and uncertainty that primes voters to believe conspiracy theories or false claims of fraud.

Read more — Are reporters doing a good job covering protests?

We have already seen the Trump administration lay the groundwork: surveillance and intimidation of protestors and physical violence against the administration’s opponents. Violence can also serve as a pretext for the crackdown itself — a rationale for “restoring law and order” that conveniently expands federal power over dissent. When Renée Good and Alex Pretti were killed in Minneapolis, video footage undercut the federal government’s claims of terrorism almost immediately, showing how far the official narrative had drifted from reality.

As communities across the country organize to ensure every eligible voter can safely participate in the midterm elections this fall, responsible coverage is not a luxury; it’s a democratic necessity. All of us should ask whether the news we’re reading or watching gives us facts we can act on, or just feelings we can’t.

There is a lot each of us can do to help protect our democracy in this critical moment. Caring for our neighbors, engaging our communities, and finding ways to use our voices are all essential. As we confront the dual challenges of political violence and authoritarian attempts to weaponize it, being clear-eyed about both the real risks of violence and the risk of its exploitation is a responsibility we all share.

Autocratic leaders around the world often use prosecution strategically — neutralizing some opponents and sending a signal to others, while preserving just enough credibility in the justice system to remain useful. Early in this term, Donald Trump and his loyalists appeared to be following this playbook. They targeted perceived opponents for technical legal violations like errors on mortgage applications, the kind of colorable but minor infractions that authoritarian regimes have long used to sideline rivals. Charges like these can be just plausible enough to keep some would-be detractors from speaking up.

That pretense is gone. In its place are ambitious displays of performative loyalty by subordinates.

Subscribe now

The DOJ’s reported investigation into Reid Hoffman’s nonprofit is the latest example. Prosecutors are investigating American Future Republic over its funding of E. Jean Carroll’s civil lawsuits against Trump — cases in which Carroll won $5 million in 2023 and $83.3 million in 2024.

The investigation has two targets. Hoffman is the primary one — a warning to major donors who might fund similar challenges to Trump. But the 82-year-old Carroll is in the crosshairs, too, in what appears to be nothing more than vengeance for her successful litigation against the president. The prosecutors’ theory centers on a 2022 deposition in which Carroll said she was receiving no outside funding. This ploy didn’t work at trial — the judge found no credibility issue and barred Trump’s lawyers from raising it, and the Second Circuit upheld that ruling on appeal. The DOJ is now trying to do what two courts already refused. Whatever the legal theory, the veneer is transparent: This is what punishment looks like when it’s no longer trying to hide.

Read more — No attorney general can deliver on Trump’s retaliation agenda

Seashells and sycophants

The second indictment of James Comey follows a similar pattern — one we outline in more detail in our tracker of this administration’s retaliatory prosecutions.

In Comey’s case, prosecutors have abandoned even a pretense of legitimacy. The DOJ is accusing the former FBI director of threatening the president’s life because he posted a photo over a year ago of seashells on a beach arranged to look like the numbers “86 47.” In the restaurant industry, “86” means to take something off the menu. Trump is the 47th president.

The DOJ is alleging that restaurant slang is a literal blueprint for violence. Comey, of course, is not a tactical threat to the president’s life; he is a retired official with a Substack. A former DOJ official dubbed this the “worst case” filed in their lifetime. Legal analysts like CNN’s Laura Coates have called it “laughable.” But seriousness isn’t the point.

Read more — Tracking retaliatory use of arrests, prosecutions, and investigations by the Trump administration

As we’ve noted before, prosecutions like these are performative acts of loyalty-signaling. Acting Attorney General Todd Blanche is auditioning to keep the role full-time by demonstrating a willingness to pursue charges irrespective of facts and law. Blanche recently made his personal priorities clear when addressing the uncertainty over who will lead the DOJ permanently. If President Trump “chooses to nominate me, that’s an honor,” Blanche told reporters. And if not? “I will say, ‘Thank you very much. I love you, sir.’”

The serious consequences of comical behavior

The instinct for visible fealty has trickled down from the courtroom to the very shoes officials wear. Senior officials like JD Vance and Marco Rubio have started wearing identical Florsheim leather wingtips because the president likes them — even if the shoes don’t actually fit. In this administration, mimicking the president’s taste is a more valuable professional asset than actually being able to walk comfortably in one’s own shoes.

While it might seem laughable, performative flattery has serious consequences. Targets of farcical cases must still hire counsel and defend themselves against criminal charges. And each act of leader-pleasing further degrades the moral and legal authority of the DOJ — and ultimately, the principle that we are a country of laws, not men.

When those wielding state power fear displeasing the president more than they fear looking ridiculous, they’re incentivized to overshoot. From cabinet officials on down, this logic can lead individuals to pursue ever-more aggressive, creative, and absurd tactics, outpacing the actual top-down directives to punish opponents and consolidate power.

Implications for the midterms

As our colleagues explained in detail in our recent Executive Override report, the White House is proceeding with its orchestrated campaign “to deceive the electorate about the security of their elections, to disrupt the electoral process, and to deny the results” of the midterms if opposition candidates win. At the same time, efforts to protect free and fair elections must account for the reality that executive actions are increasingly driven by subordinates devising new ways to flatter Trump and benefit themselves professionally — to secure a nomination, or simply keep their current job.

Read more — How the Trump administration plans to interfere with the 2026 elections, and what you can do about it

The case of Thomas Albus is instructive. Albus, a Missouri prosecutor whose earlier work in the U.S. Attorney’s Office didn’t involve election law, was tapped by then-Attorney General Pam Bondi and given special authority to handle election-related cases nationwide. Albus understood the assignment. He coordinated with Ed Martin and White House lawyer Kurt Olsen — both of whom had worked to overturn the 2020 election — and then authorized the FBI seizure of roughly 700 boxes of election material from Fulton County, Georgia, far outside his usual jurisdiction. The raid — which legal experts described as a significant escalation in the breaking of democratic norms — wasn’t traced to a Trump directive; it emerged from a network of loyalists who had built careers anticipating what he wanted.

Self-starting sycophants could very well play a starring role this fall, as ambitious officials seek out opportunities to ingratiate themselves by tilting the playing field in favor of Trump’s preferred candidates — potentially without waiting for an explicit directive from the White House. Now that “bootlicking” is the currency of the realm, one of the serious challenges we encounter this fall could be as simple as a prosecutor looking for a promotion.

Laughter isn’t the only medicine

When a federal indictment is based on a message spelled out in seashells, a reasonable response is laughter. Humor is also a bedrock tactic in resisting those who demand to be taken seriously while acting absurdly. But as satisfying as it is to call out the absurdity, we can’t let the cartoonish nature of these actions distract us from their underlying danger — or from the opening they create.

What makes these prosecutions so hard to stop is that subordinates often don’t need to be told. They know what Trump expects, and they know what happens to those who don’t deliver it. The result is a race to the bottom — ever more ridiculous and sycophantic displays of performative loyalty. But sycophantic prosecutions are, by definition, bad prosecutions. And bad prosecutions are vulnerable ones.

No single institution can fix this. But there are meaningful steps we can all take — and the very absurdity of these cases gives us more tools than we’d otherwise have:

• Criminal defendants can demand fee awards from the DOJ under the Hyde Amendment when charges are “vexatious, frivolous, or in bad faith.” They can also pursue civil claims for malicious and vindictive prosecution to keep the spotlight on misconduct and seek accountability.

• Federal judges can set aside the “presumption of regularity,” as DOJ and other federal agencies no longer deserve a default rule that assumes they’ve acted in good faith and in a lawful manner. Judges can demand greater transparency into dubious prosecutorial behavior, granting tailored discovery requests, reviewing grand jury transcripts, and authorizing public disclosure when necessary and appropriate.

• Federal district courts can update their local rules to ensure adequate oversight of DOJ’s failed efforts to secure indictments before grand juries.

• Congress can exercise oversight of the DOJ and other federal agencies, demanding transparency into dubious executive actions and imposing clear limitations on the use of appropriated funds. The Senate can reject nominees who broke the law or otherwise disqualified themselves in the pursuit of presidential approval.

• State bar officials can strictly and promptly enforce their ethical requirements against federal prosecutors and other attorneys in the executive branch when presented with credible allegations of misconduct.

• State legislatures can authorize private lawsuits against federal agents who violate their constitutional rights — including officials who retaliate against individuals for criticizing the president or otherwise engaging in protected First Amendment activity. Over the last few weeks, New York and several other states have passed such laws.

• Individuals can follow and share the work of journalists and watchdog organizations tracking prosecutorial misconduct. Public pressure and reputational cost can be real constraints on bad-faith prosecutors — the Comey and Carroll cases are generating pushback partly because coverage has made the absurdity visible. Readers who amplify that coverage are participating in the accountability ecosystem.

Governance by flattery is no way to run a democracy. While federal officials bend over backwards to impress an audience of one, the rest of us must work even harder to protect our elections and the rule of law.

The Supreme Court’s decision to gut what remained of the Voting Rights Act is the latest blow to our democracy; states in the South immediately began mobilizing to redraw congressional and state legislative districts in a way that further dilutes the voting power of Black and Latino voters.

Subscribe now

The situation for minority voters was dire to begin with. Latinos, for example, are dramatically underrepresented in Congress, occupying only 11 percent of its seats despite making up 20 percent of the population. Our voting power has been diluted by extreme gerrymandering for decades, which will likely grow more severe this year. In Texas, for example, where nearly 40 percent of the population is Latino, minority voters have long been underrepresented in the House delegation. The new gerrymandered voting map in Texas was specifically designed to eliminate Latino and Black-held seats; we will see more of this as a result of the Callais v. Louisiana decision.

Because this year’s midterm elections are vital to the fate of our democracy, with their potential to check the president’s authoritarian agenda, the Administration’s opponents are responding to these dynamics with gerrymandering efforts of their own in states controlled by Democrats. While this is an understandable impulse, given the high stakes, extreme gerrymandering is terrible for the health of our democracy in the long run. Even before the current rush to redraw districts, only 4 percent of congressional races this year were judged to be competitive overall.

Uncompetitive elections are a sign of a broken system that affects all Americans and contributes to our disillusionment with both major parties, as well as with our seeming lack of ability to do anything about our nation’s political gridlock. We are living through a year replete with evidence that our winner-take-all election system incentivizes the creation of uncompetitive elections that further political polarization. This has contributed mightily to the horrifying excesses of the Trump era, while Congress seems powerless to check even the President’s most egregious actions.

But even if we were somehow to prevent the further gerrymandering that dilutes minority voting power, it is not clear that even districts where Latinos and Black Americans are the majority of voters are the best way to ensure that our voices are heard fairly. In states with smaller but still significant Latino populations, like Colorado or Georgia, for example, the number of districts we control is likely to remain small, even without gerrymandering.

Beyond that, the idea that majority-minority districts will concentrate power among people whose views are similar because of their shared ethnicity has never held up in a community as diverse as ours. Just ask a Latino Republican in California or a Democrat in Florida.

We can do so much better than this. A growing body of research strongly suggests that Americans of all types, including Latinos, would benefit from a move to a fairer system for all voters, one that aligns voters’ actual preferences more closely with our electoral outcomes, and where every vote matters. A different electoral system — one used by most other mature democracies around the world — would ensure this: under Proportional Representation, multi-winner districts are used to allocate votes proportionally so that our elected representatives reflect the actual preferences of voters.

Read more — Proportional representation, explained

In a single-member district like the ones we have now, 51 percent of voters can elect one candidate, and the other 49 percent get nothing. But if four single-member districts collapse into one broader district that elects four winners, a minority population that makes up 25 percent of the district can elect one of the candidates. The 50 percent majority can still elect two candidates in proportion to their numbers, too. This means that voters are represented in proportion to their preferences regardless of whether the district lines are drawn to favor them.

In fact, gerrymandering tends to be far less of an issue in democracies with proportional representation. Because parties win seats according to their share of the vote, it typically doesn’t matter much how voters get shuffled around — wherever they are, their electoral power generally remains the same. And because votes are more likely to matter everywhere, parties and candidates are incentivized to broadly compete for those votes — including among voters of color. The system could even open up our two-party gridlock by allowing new parties to compete, reflecting a multipartyism in line with the diversity of American political perspectives.

Latinos have never waited for permission to shape this country’s future. We’ve built communities, forged coalitions, and fought for our voice in a system that was rarely if ever designed with us in mind. This moment, as dark as it feels, is not new. But we now have a growing movement of Americans from a diversity of backgrounds who want the same thing we do: elections that actually capture the will of the voters, representatives who have to earn every vote, officeholders who reflect the diversity of our communities, and a politics capable of solving very real problems rather than manufacturing endless conflict.

We should not continue to accept a system that dilutes our power, rewards extremism, and leaves the majority of Latinos and Americans feeling like our vote doesn’t matter. The choice before us now isn’t just between the two major parties: it’s between the broken system we inherited and a fairer one we can build together. For Latino voters and organizations, building a politics where our voices are heard and our votes counted is the next frontier of the fight for our representation and our rights. In this moment of peril for our democracy, it’s a fight we should take on.

This piece was originally published in Democracy Journal.

The second Trump administration has set about dismantling the institutions of American science.

Over the past 16 months, the administration has eliminated thousands of federal scientific positions, imposed sweeping funding cuts at the National Institutes of Health (NIH) and National Science Foundation (NSF), dismantled the Centers for Disease Control and Prevention’s (CDC) vaccine advisory committee, and worked systematically to manufacture political justification for abandoning the evidence that underlies public health and environmental policy.

Subscribe now

These attacks are not random. They are interconnected parts of a nationwide strategy to erode shared truth, manufacture doubt, and teach the public to distrust what it once took for granted.

Science has been a testing ground. If you can make people doubt peer-reviewed research, vaccines, and climate data, you can make them doubt anything. An electorate that doubts everything is an electorate that can be told anything.

So as the 2026 elections approach and the administration escalates its efforts to deceive voters and deepen that distrust, it would be fair to assume the scientific community might stay quiet based on its initial response in 2025. When the first attacks came, the impact was profound. Many scientists were paralyzed by fear of losing funding, fear of political retaliation, fear that speaking out would cost them everything they had spent their careers building. Some self-censored, avoided controversy, and waited for the storm to pass.

Read more — Courage is (mildly) contagious

But something has started to shift. Scientists are fighting back. A resistance that began as individual acts of courage is now taking collective shape, and what has emerged over the past several months is beginning to look like the early architecture of a movement.

Inside out and outside in

Inside federal agencies, scientists and career workers have faced these pressures most directly. The Federal Workers Alliance for Democracy (FWAD) has built a growing coalition of workers and allies mobilizing the federal workforce to refuse compliance with dangerous and illegal orders, building networks across every federal agency in every state. Aisha Coffey, FWAD’s communications director, explains:

The thing about messing with career civil servants is that those are the same people this administration will rely on to carry out dangerous, illegal orders. If federal workers refuse to comply, we take an extraordinary amount of power away from a government that’s trying to do us harm.

It’s a powerful strategy. And to add to it, the resistance isn’t only coming from inside. Scientists are also building power from the outside.

In June 2025, Health and Human Services Secretary Robert F. Kennedy Jr. dismissed all 17 members of the CDC’s Advisory Committee on Immunization Practices (the body that has guided U.S. vaccine policy since 1964) and replaced them with a group that included vaccine skeptics and anti-vaccine advocates, some lacking relevant expertise.

The dismissed scientists did not simply step aside. They formed the Immunization Scientific Advisory Collaborative (ISAC) and published independent, evidence-based evaluations of the reconstituted committee’s proceedings. When the committee voted to end the decades-old recommendation that every newborn receive a hepatitis B vaccine at birth (despite no new safety data and no new evidence that the shots don’t work), ISAC documented what that process looked like. Together, they created a public record of what evidence-based vaccine policy requires and clearly told the public what had been abandoned.

Young scientists are adding their voices to the chorus

While many of these examples involve established scientists and career workers, early-career researchers have stepped up too. Without tenure, established reputations, or institutional protection, they have the most to lose by speaking out, but this isn’t stopping them. Graduate students turned hallway conversations into the Scientist Network for Advancing Policy (SNAP), now a nationwide coalition of more than 150 active members across 30 organizations.

SNAP’s McClintock Letters initiative coordinated more than 600 scientists to write op-eds for their hometown newspapers (not elite national outlets, but local papers) in the communities they grew up in. They have already published more than 200 pieces across 45 states. As JP Flores, speaking on behalf of SNAP, explains:

We wanted to get back in touch with the people we grew up with, the communities that shaped us, and the people who, in many cases, we became scientists for in the first place. We hoped our peers would join us and were amazed by how many did. If scientists want to continue to earn the support and the funding we’ve received from the American public, we have to not only communicate with them but also recognize their vital role as partners and stakeholders in our work.

The scientific community has been criticized, often fairly, for operating at a remove from the public. These early-career scientists are working to close that gap at precisely the moment it matters most.

Healthcare workers were crucial to resistance in Minnesota

Nowhere has that urgency been clearer than in health care. When the Trump administration designated hospitals as fair game for immigration enforcement, the Minnesota Nurses Association doubled down on their efforts to educate members about patient rights, what to say to federal agents, and how to protect immigrant patients, including Minnesota’s large Somali and Hmong communities. They developed badge buddies for nurses (printed cards with instructions and a QR code for resources) and built partnerships with immigrant rights groups, faith organizations, and legal aid groups.

Then, on January 24, 2026, federal agents shot and killed Alex Pretti, a 37-year-old ICU nurse at the Minneapolis VA. He was off duty, observing protests against Operation Metro Surge. That day, he stepped between a federal agent and a woman who had been pushed to the ground. Alex had spent years helping care for patients in the ICU and was doing the same for a neighbor. Federal officials said he had brandished a weapon. The video tells a different story.

Pretti’s death galvanized nurses nationwide. National Nurses United declared that “ICE messed with the wrong profession” and organized a nationwide week of action: candlelight vigils at hospitals and VA facilities from Minnesota to California to New York, with explicit demands that Congress defund ICE or face electoral consequences.

Year after year, Gallup finds Americans rank nurses the most trusted profession in the country. When a profession like that organizes, it brings the power of everyone it has spent decades caring for.

Where to plug in

These examples are far from comprehensive. Organizations like HealthBegins have been training frontline clinicians to treat pro-democracy work as an extension of their professional identity, not a betrayal of it. Stand Up for Science returned to the streets in March 2026 with rallies in 46 locations nationwide. New cross-sector coalitions are forming in states across the country. Momentum is building as scientists get in formation ahead of November 2026.

One example of this growing momentum: On June 3, the Union of Concerned Scientists is launching Science Rising, an initiative aimed at mobilizing the science community for sustained engagement, building the political infrastructure to defend science and democratic accountability through the 2026 elections and beyond. As Gretchen Goldman, president and CEO of the Union of Concerned Scientists, says:

If ever there was a time for scientists to step into our power, it is now. Science Rising recognizes the gravity of the threats to science and democracy at this moment and harnesses the power of scientists and science supporters to fight back.

The launch webinar on June 3 (3-4 p.m. ET) will feature three concrete actions: signing up for SMS alerts for rapid mobilization, raising awareness in your own networks, and joining the Science Rising Action Corps for tools and opportunities focused on congressional accountability throughout the summer and fall.

If you are a scientist (or just someone who believes science should serve the public good) RSVP for the launch here. If you work in science or technical fields specifically, there is a scientist-specific form here.

Over a year ago, the fear was felt at the individual level. A scientist’s grant withdrawn, or their name added to a list; a career built carefully over years at risk of being lost overnight. The administration bet that the fear of thousands of scientists, isolated, weighing the costs of speaking up, would add up to compliance. But the math has changed as the scientific community is proving them wrong by remembering it is one.

Former Attorney General Pam Bondi did everything President Trump asked her to do with gusto. Trump’s primary objective for his Attorney General (AG) was to use federal law enforcement to punish his perceived enemies and opponents, and Bondi tried hard — and largely failed — to do just that.

Subscribe now

She indicted former FBI Director James Comey, New York AG Letitia James, and a member of Congress conducting ICE oversight. She launched investigations of Federal Reserve Chair Jerome Powell, former impeachment prosecutor and now senator Adam Schiff, former CIA Director John Brennan, and Democratic members of Congress who restated the basic principle that military personnel can disobey orders they know are unlawful. The list goes on. After most of these cases were dismissed by federal district court judges, rejected by grand juries in unprecedented numbers, or languished in the investigative phase, Trump fired Bondi, reportedly because he believes she wasn’t strong enough in her resolve to pursue his vendettas or smart enough to win the cases.

It’s hardly surprising that this is Trump’s view of one of the many women he has scapegoated for his failures. But as a factual matter, it’s wrong. Bondi didn’t suffer from reticence or incompetence. She failed thanks to the American people — the grand jurors, trial jurors, and judges who refused to let meritless cases proceed. These are guardrails the president cannot fire, reassign, or intimidate into submission.

Read more — A guide to understanding the federal grand jury

Todd Blanche, Trump’s personal attorney and now Acting Attorney General, will run into the same wall. So will whoever comes next. This is good news for democracy, but it comes under a dark cloud. While Trump’s retribution campaign is failing, a Justice Department that should be protecting ordinary Americans has abandoned that mission in futile pursuit of the president’s vendettas.

Bondi was willing, just unsuccessful

Protect Democracy has documented a sustained pattern by Bondi of investigating and prosecuting Trump’s political opponents and critics. The cases have typically involved public smearing of the subjects in violation of the Department’s Principles of Federal Prosecution and Constitutional due process, direct interference from the president and other political appointees in charging decisions, overriding the judgment of career professionals, and an outright lack of evidence of any criminal wrongdoing.

None of this seemingly gave Bondi any pause. And if she ever seriously entertained any of the concerns her DOJ subordinates communicated to her about losing meritless cases in court, she proved herself willing to brush them aside when the president told her she should.

Read more — Tracking retaliatory use of arrests, prosecutions, and investigations by the Trump administration

Bondi was stopped by the legal system’s guardrails, not her own limitations

Of course, Bondi did lose a lot of cases in court. And she lost just as often before she could even get a case started. But a more skillful or aggressive lawyer would not have fared better. The Comey and James cases were both widely criticized as lacking evidence of the commission of any crimes and were dismissed because the district court held that the U.S. Attorney who obtained the indictments was unlawfully appointed. The Justice Department appealed the dismissals and has repeatedly attempted — and failed — to re-indict Ms. James.

This is of a piece with the role grand jurors have played in thwarting the Justice Department’s retaliation agenda by rejecting an unprecedented number of indictments. After years in which grand juries issued no-bills — refusals to approve charges — in single-digit numbers out of more than 150,000 cases pursued, grand juries during the second Trump administration have rejected charges in dozens of cases all over the country. Trial jurors have also been a significant roadblock, acquitting people facing dubious charges arising out of protests against the Trump administration’s mass deportation campaign. In Washington, D.C., for example, a jury acquitted a former federal employee who was charged with misdemeanor assault (downgraded from a felony after a grand jury refused to indict) for throwing a Subway sandwich at a Customs and Border Patrol (CBP) agent. After the trial, one juror told the media that people in the courtroom had a hard time keeping a “straight face” listening to the government’s evidence. “I thought we’d be out of there quickly,” the juror said of deliberations, because “[t]his case had no grounding.”

This reality is starting to sink in with some Trump appointees, who are closing cases and dropping charges rather than continuing to be humiliated in court. To take one recent example, after a grand jury unanimously refused to indict the Democratic members of Congress who spoke out about disobeying unlawful military orders, D.C. U.S. Attorney Jeanine Pirro ended the investigation instead of making another attempt to secure an indictment.

That said, the danger isn’t over. Acting AG Blanche has already announced a legally dubious indictment of the Southern Poverty Law Center — apparently in retaliation for the organization’s designation of certain Trump allies as “hate groups.” And he followed a few days later by announcing yet another indictment of Comey, this time for allegedly threatening to assassinate Trump via a message communicated in seashells and posted on social media. Legal experts immediately pronounced the indictment deficient on First Amendment grounds.

The hardships these indictments cause to the defendants should not be underestimated. But no AG, however committed or clever, is likely to have much success getting factually empty cases past American juries. And the Trump administration cannot secure indictments and win convictions without the cooperation of the very citizens it’s trying to intimidate and punish. Notably, the government has no right to appeal grand jury no-bills (although it can try again) or trial jury acquittals; here, the people have the last word. The administration of criminal justice is one realm where the American people have significant power over an autocratic president, are using it to great effect, and are showing no sign of changing course.

Bondi seriously damaged the Justice Department’s ability to engage in legitimate law enforcement

Yet, even as ordinary Americans are succeeding in blocking Trump’s illegal retaliation agenda in court, they remain its primary victims outside of it. As she was pursuing — and losing — retaliatory and vindictive prosecutions, Bondi decimated the ranks of competent career professionals in both the Justice Department and the FBI and upended the Department’s legitimate law enforcement mission. During her tenure, more than 16,000 employees, including experienced prosecutors and FBI agents, were fired or resigned. The Justice Department’s Public Integrity Section — the unit that specialized in public corruption and white-collar prosecutions — has been effectively shut down. The Criminal Section of the Civil Rights Division, which prosecuted law enforcement misconduct and hate crimes cases, has also been gutted.

This loss of law enforcement expertise and capacity, together with the Trump administration’s focus on its political agenda of immigration crackdowns and political retaliation, has led to a precipitous decline in enforcing drug, white collar, public corruption, and national security crimes, among others. ProPublica reported that the Trump Justice Department “closed more than 23,000 criminal cases in the first six months” of the administration, a marked increase over previous administrations. Prosecutions in areas from drug trafficking, government procurement fraud, terrorism, and national security have all declined sharply.

None of this is consistent with keeping the American people safe from violent crime or protecting them from fraud, corruption, and other threats to their financial security and environment. And it is not a recipe for another Attorney General to win more cases of any kind. Instead, by continuing to pursue baseless and retaliatory cases, Trump’s Attorney General will inevitably force out more of the ethical prosecutors who remain. Indeed, that has already happened during Blanche’s tenure, as a top career prosecutor in Miami was removed from overseeing the investigation of former CIA Director John Brennan after she reportedly expressed doubts about its legal viability.

A reason for optimism, and a warning

Whether the next Attorney General remains Todd Blanche or Trump nominates another loyalist to continue pursuing his retribution campaign, the American people and other institutional actors will almost certainly keep blocking the Justice Department’s illegal actions. This should be a point of inspiration and real cause for hope that our democracy can survive — one of the best we have been given in the last two years. But every day that the Justice Department remains an instrument of presidential vengeance, the harder it will be to rebuild an agency worthy of the very people who — throughout this federal dereliction of duty — have done so much themselves to protect justice.

You may have noticed: There are a lot fewer news stories about the Trump administration dramatically killing federal programs these days. From the original memo freezing funds to DOGE cancelling contracts and paying federal employees not to work — there’s simply less noise and drama over spending abuses in 2026 than there was in 2025.

But that doesn’t mean the abusive cuts have stopped. Instead, the Office of Management and Budget (OMB) has often shifted to killing programs quietly. If the DOGE cuts were like putting federal programs up for public executions, these days the abuses happen more like knives in a back alley.

Subscribe now

What do those back-alley knifings look like?

Take the Community Development Financial Institutions (CDFI) Program. This program invests in community financial institutions to provide loans, and financing to areas in need of capital to support the development of affordable housing, and investments in farmers and small businesses, among others. Advocates estimate that every dollar in federal money leverages $8 in private capital that goes toward investments like more affordable housing.

The program is backed by a bipartisan caucus of legislators, and a coalition of groups that include a range of different types of organizations from banks to community lenders to housing advocates.

But the most important person when it comes to spending money in the Trump administration is OMB Director Russell Vought — and Vought hates this program.

That’s why OMB is killing it.

Since Vought took over OMB, not a penny of the $578 million of the 2025 or 2026 appropriations Congress passed for the CDFIs has been made available for the Treasury Department to spend.

To do so, OMB is using an obscure part of the budget process called apportionments to murder the program through small cuts — though rather than a thousand cuts, in this case it only takes five.

How to kill a program in five cuts

Each year after the president proposes a budget and Congress passes spending bills, it’s OMB’s job to apportion the funds, issuing documents for each account telling agencies how much money they can spend and how fast.

Until they get an apportionment, an agency generally cannot spend funds — even if they know the money is coming, even if they know Congress passed a law sending it to them, they still have to wait for an apportionment.

And that is how OMB made its first cut — it left the agency waiting. In fact the agency is still waiting — almost a year later.

In 2025 Congress passed a law providing $324 million for the CDFI program, $35 million for Treasury to spend in fiscal year 2025 to administer the program, and $289 million for Treasury to spend on funds for community banks and developers to actually help people do stuff.

In the apportionment process that normally happens, once Congress passes the bill, OMB apportions the funds, making the $35 million available for Treasury to spend in the current fiscal year, and transferring the remaining $289 million to an account that lasts for two years. This gives the agency a little more time and flexibility to spend those hundreds of millions of dollars responsibly.

What happened in 2025 was different. OMB apportioned the $35 million for program administration. And it transferred the $289 million to an account from which the money can be spent for two years, 2025 and 2026. But then OMB never apportioned the $289 million in program funds. To this day they are just parked in that two year account, aging in place.

Last year, advocates attempted over and over again to persuade the administration to release the money. They sent this letter about the benefits of the program and the communities it serves. They organized this letter from over a hundred Republican members of the House and Senate.

But still no apportionment of the 2025 funds. Advocates have continued to press. The Treasury Secretary recently stated he would spend the CDFI money if OMB released it. And given the amount of time it takes for the agency to complete its review of applications, if OMB doesn’t release the money soon, the 2025 funds will expire at the end of September without being spent.

Not only has OMB refused to release the 2025 funds, it has taken steps to hold back more.

There are at least three other accounts that have some funds that have gone to support CDFIs and their work, all three of those accounts collect proceeds from various mandatory funding streams. The money in those accounts is available indefinitely, rolling over year after year, and the lion’s share of those funds have been apportioned in the past for the CDFI program.

But over the last six months, these three accounts had almost all of their funds apportioned into Category C, and not a penny was made available to Treasury for the CDFI program.

Technically Category C is a way to apportion money for future fiscal years. You can see how it would come in handy if Congress passed a big pot of money for something like a moonshot research program, OMB can use Category C to structure the spending. So for example if an agency was still spinning up an office to manage the funds, you could give them time to get ready before sending them billions of dollars to manage and spend.

But over the last year OMB has started using Category C for more than money management. OMB has started using Category C to hold back funds that Congress told them to spend. And because funds apportioned in Category C cannot be spent in the current fiscal year — those apportionments effectively act as a cut.

Category C is a sneaky way for OMB to unilaterally make cuts it could not get Congress to agree to in the law.

And that is how OMB made three more cuts to the CDFI program — by preventing even a single penny of the CDFI Program account, the Emergency Capital Investment Fund or the Capital Magnet Fund from being spent on the CDFI program this year — accounts that used to provide tens or hundreds of millions for the program.

In the last few weeks, advocates have been raising the alarm that between the 2025 money it impounded and the rollover funds it froze in Category C, OMB is holding up more than a billion dollars that could be used for CDFIs.

Two weeks ago, OMB responded by making a fifth cut — pulling the same move with the fiscal year 2026 appropriations for the CDFI program.

Just a few months ago Congress passed a law on a bipartisan basis providing $324 million for the CDFI program, $35 million for administration and $289 million for the program itself.

And once again, OMB apportioned the administrative funds, but transferred the $289 million to another account, an account for which there hasn’t even been a public financial report yet, let alone an apportionment.

To the extent that anyone thought OMB would have more respect for full year appropriations bills passed on a bipartisan basis, this incident ought to put such hopes to rest. If OMB wants to kill your program, shame over breaking the law won’t stop it from trying.

Spending abuses are illegal, regardless of how sneaky

But hey, you’re thinking, this sounds illegal.

Yes! It is.

And sure, the CDFIs could sue. But that’s going to take time. And they’ll have to convince a court that it’s an emergency to force OMB to apportion it now — even though the 2025 money isn’t going to expire for six months — so that Treasury has time to take in applications and review them and decide which CDFIs get what funds. And if they can’t convince a court to force OMB to get the money out sooner than later, some or all of that $289 million will expire.

In the meantime, the CDFI program, much like someone who has been stabbed five times in a back alley, is vulnerable and at risk for more attacks.

While CDFIs fight to claw back this money either in federal court or the court of public opinion, OMB can propose a rescission of these funds that would send this money back to the Treasury. Or a legislator looking for an offset to fund their pet project might use some or all of the billion bucks sitting in these accounts to pay for something else. And if either of those things happens it won’t matter if the CDFIs win their case, because the money will already be gone.

So, you ask, is OMB really just going to get away with it? Is this like the perfect money murder?

Maybe, though there is one way they might get caught.

When Congress passes spending laws it attaches all kinds of conditions. A few of those conditions are called ‘general provisions’ and they apply to all the appropriations across all the agencies.

For years, one of these general provisions has prevented agencies from using any funds to implement cuts in the president’s budget unless or until Congress agrees to those cuts in an appropriations law.

And right now OMB appears to be violating that condition — a condition which applies to OMB’s own funding. By holding back the money for a program Congress funded but the president wanted to eliminate in 2026 OMB is basically doing what Congress says it cannot use funds to do — going rogue and unilaterally implementing the president’s budget.

And the law that enforces these kinds of conditions is a scary one — the Antideficiency Act — it imposes personal liability on the officials that violate it.

Here’s the kicker — the statute of limitations for this law is five years. So what can CDFIs and Congress do?

They can put OMB officials on notice that they may be liable, and they can keep a record of who has been responsible for what. Because if a future DOJ can prove an official violated the Antideficiency Act knowingly and willfully, they won’t just be liable for administrative penalties.

For the crime of money murder, they could face up to two years in jail.

Authoritarians often thrive on chaos. Since first taking office in 2017, Donald Trump has defined almost every news cycle, and he’s often created so much news that it can be hard to distinguish signal from noise.

The chaos is the strategy.

Even for those of us whose work involves following government and democracy, it can be difficult to keep up with every development. When people can’t keep track of what’s happening and what it means for the things they care about, they’re more likely to throw up their hands in defeat. Modern authoritarians don’t want us to be able to understand the entirety of what’s happening to our democracy. That’s why they disguise their assaults in a flurry of actions, scandals, and outrages.

Subscribe now

That misdirection is part of a larger effort. Authoritarianism works through a cycle of fear and acquiescence. One by one, would-be dictators convince everyone who would stand in their way to voluntarily surrender. To obey in advance.

In this effort, autocrats take advantage of a fundamental aspect of human psychology: When we feel uncertain, we feel afraid. When we feel overwhelmed, we’re more likely to retreat — to back away and try to protect ourselves from dangers unseen or unknown. By flooding the zone, Donald Trump and his supporters are trying to break down our rational faculties and scare us into a sense of panic and confusion.

Over the past year and a half, we’ve seen this pattern play out across law firms, universities, media companies, and more. When the attacks feel overwhelming, targets are much more likely to simply crumple.

That’s where our new tool comes in.

Because no matter how bad things may be, objectively, clarity is always better than confusion. Understanding the specific contours of threats makes it much easier to get our heads around what needs to be done. And to steel ourselves against coming attacks.

The Authoritarian Action Watch is designed with the explicit goal of helping every American better understand what recent news developments mean for American democracy. In that clarity, they can find their own sense of resolve.

A new tool for discerning signal from noise

The Authoritarian Action Watch is designed to answer one question: How are we doing, as a democracy? To do that, it offers users an overall snapshot that works a little like a weather report. It’s not based on an absolute scale (the temperature can always get hotter), but a directional one that tells you how much worse or better things have gotten over the past few weeks.

To generate that snapshot, we rely on the Authoritarian Playbook, which describes the seven tactics authoritarians rely on all over the world to consolidate and hold power. For each of these tactics, Protect Democracy’s experts analyze events in the news, and determine what they mean for the trajectory of American democracy.

Is this good news, bad news, or, as is often the case, a mix of both?

This tool is designed to illustrate the way our democracy is moving at a given moment, but it also seeks to explain why the Trump administration is making some of the moves it’s making. You might learn, for example, that deployments of ICE and CBP agents to cities across the country are not only about immigration enforcement. Instead, they’re part of a larger strategy aimed at quashing dissent, stoking violence, and ultimately making Americans feel less free to exercise their rights, including their right to choose their leaders.

Even as we analyze the actions of the Trump administration and what they mean for our democracy, we also want the Authoritarian Action Watch to offer signals about all the ways the democracy movement is pushing back. Over the first year of Trump’s second term in office, it often felt like authoritarianism was ascendant, and there was very little standing in its way. What we saw over the course of that first year, though, was that regular Americans are ready to take a stand to protect the ideals this country has always stood for, albeit imperfectly.

So even as we track domestic deployments, acts of executive entrenchment, and attempts to capture the media, the Authoritarian Action Watch will also highlight the actions of people with far less power than the president and his administration who are nonetheless standing up to defend democracy. This tool will capture the parents who are part of school drop-off caravans in Minneapolis, the state legislators who are bucking their own party on crucial votes, and the grand juries who are standing in the way of weaponized indictments. These are crucial reminders that this fight is not already lost.

We believe the Authoritarian Action Watch can be a useful tool for helping you understand what all of us are living through. The coming months and years will be some of the most trying for our democracy in decades. This tool should help us understand every twist and turn, and then, help us do something about it.

Explore the Authoritarian Action Watch

Recent observers of the Supreme Court are coming to terms with an ugly truth: a majority of the justices seem to be re-considering the legality of Section 2 of the Voting Rights Act (VRA). In the pending landmark case Louisiana v. Callais, the Court may decide that this key safeguard against race-based voting discrimination can no longer be enforced.

The stakes of this impending decision are even broader than they appear.

Subscribe now

What many court-watchers may not fully realize is that further weakening Section 2 of the VRA not only undermines voting rights, but also debilitates a vital power-sharing system: compelling political majorities to include political minorities. To protect our democracy from this looming winner-take-all threat, we must consider a structural alternative — proportional representation.

End of an era?

For decades, Section 2 of the VRA protected voters of color from having their voting power diluted. It helped create “majority-minority” districts, ensuring that Black voters — who historically make up 20% to 40% of most Southern states — were no longer perpetually shut out of government by white majorities.

Callais threatens to dismantle this framework. The lawsuit centers on Louisiana, where Black residents make up one-third of the population, yet the state drew only one majority-Black congressional district out of seven. During arguments, several justices signaled that the majority-minority districts that have served as the primary remedy for Section 2 violations are no longer a valid solution to racial discrimination in voting.

However, as articulated in a recent Yale Law Journal essay, Callais Confusion, Power-Sharing, and the Inevitability of Proportional Representation, dismantling Section 2 destroys a key mechanism for ensuring that our democracy accurately reflects the racial and political diversity of our society.

Since the founding of the republic, constitutional scholars have warned about the threat that unchecked majorities pose to our political rights. For 60 years, the VRA has served as one of the few electoral mechanisms to protect against the “tyranny of the majority” and prevent large political factions from invading the rights of other citizens.

If Section 2 is struck down, it will be the end of an era. One in which the VRA was a key tool to ensure not just the right to vote, but the right for your vote to matter. Our efforts to ensure that every vote is meaningful reflect what voting rights scholar Lani Guinier called “second-generation fights.”1

The first generation fight got rid of poll taxes and literacy tests so that Black voters could finally cast a vote. The second generation wins made those votes count.

The proportional solution

If the Court continues dismantling our established voting rights regime, how do we ensure that political minorities still have a seat at the table? We look to a solution already used by 20 of the world’s 24 presidential democracies: proportional representation.

In short, proportional electoral systems ensure that the share of seats a party gets is aligned with its share of votes. So rather than 51% of voters get 100% of the power and 49% get nothing, all coalitions of voters (provided their group meets a minimum threshold) receive seats proportional to their votes.

We can achieve this by replacing our existing single-seat, winner-take-all system with multi-seat proportional districts and ballot structures that allow voters to pool their voting strength. So rather than 4 districts each electing one member of the majority, one district elects 4 representatives. If a minority of voters vote together and constitute 25% of a district, they elect one representative. If another group constitutes 50%, they elect two. Everyone’s vote counts equally.

Proportional representation is a great system in its own right: a large body of research demonstrates its effectiveness, especially its ability to better represent the full demographic and ideological diversity of society.2

Proportional systems also have a unique advantage when applied to the U.S. context: It is possible to achieve fair representation for voters of color without race-conscious districting. In fact, proportional representation makes gerrymandering functionally impossible if enough seats are elected per district. That also means that the practice of drawing district lines around racial groups is not necessary to secure representation for racial and ethnic minorities.

Because proportional representation is a race-neutral electoral system on its face, it doesn’t require the government to choose in advance who will get representation. This means it could withstand scrutiny from the Court’s majority, who have grown increasingly averse to race-conscious policies.

We need a new path forward

Proportional representation is not a consolation prize for the demise of the Voting Rights Act. The VRA was implemented after a long, bloody fight for recognition of full humanity and citizenship by and for people of color in the United States, and it served as a powerful tool to enforce those rights for the last 60 years.

What will come next must be informed by the VRA’s history and its powerful defense of rights for all Americans, especially those long excluded from power.

Proportional representation should be part of the answer. In the tradition of many generations of civil rights reformers, we must continue their tireless work and begin anew to secure democracy for all.

Multiple American presidents have at some point discovered the same temptation: the strategic industry too important to national security to leave in private hands. The modern version of this story has a three-act structure that begins with Harry Truman, reaches a surprising climax with Donald Trump, and may be approaching its denouement with artificial intelligence.

The arc traces a question long disputed: What happens when a president decides a private industry is too vital to national security to operate beyond his control? Truman tried brute force with the steel industry and was told he couldn’t. Trump found cleverer mechanisms and, so far, has succeeded. But AI may represent something new — an industry where the legal tools available to the executive branch are real and formidable, and yet where exercising them might not produce the desired outcome — and may reveal something profound about where future power lies.

Subscribe now

In short, AI is not like a steel or munitions industry. It’s not just capital, infrastructure, and processes that can be captured and taken over by the government. It’s not even just companies like Anthropic or others. Instead, the power at the center of the AI industry today is a small number of leading AI scientists.

If the Trump administration tries to take over their employer by force and coerce them into maintaining a product for uses they fundamentally oppose, those researchers are likely to simply walk away or find other ways to frustrate the government’s goals. Unlike a steel mill, they cannot be forcibly commandeered into the U.S. national security apparatus. This means the present contest between the Defense Department and Anthropic has contours unlike prior showdowns between the president and industry.

When Harry Truman seized the steel industry

In April 1952, with a steelworkers’ strike threatening Korean War production, Truman issued an executive order seizing the nation’s steel mills in the interest of national defense. Two months later, the Supreme Court struck the order down. In Youngstown Sheet & Tube Co. v. Sawyer, Justice Jackson’s concurrence established the enduring framework: Presidential power is at its “lowest ebb” when the president acts against the will of Congress, which had previously considered and rejected seizure authority. The case seemed to establish a firm principle: The president cannot commandeer private industry without congressional authorization, even when national security is genuinely at stake.

But notice what Youngstown quietly assumed. The constraint on Truman was juridical, not operational. No one argued the government lacked the practical capacity to run steel mills — in fact, it had done so during World War II. If the law had been on Truman’s side, the mills would have been seized and run. For 70 years, that assumption held: The only meaningful check on presidential power over industry in the context of national security was the legal one.

U.S. Steel under Biden and Trump

When Japan’s Nippon Steel bid $14.9 billion for U.S. Steel in late 2023, both Republicans and Democrats opposed the deal. In the final days of his administration, President Biden blocked the deal, citing national security concerns. During the 2024 presidential campaign, Trump had vowed to do the same, but after being sworn in, he reversed course — but with a twist. He ultimately approved Nippon’s investment in U.S. Steel as part of a deal that gave the U.S. government a more extraordinary level of control over a private corporation than almost any president had sought or achieved since Truman tried and failed.

The restructured deal for U.S. Steel’s purchase, finalized in June 2025, required Nippon to make $11 billion in new investment, install an American CEO and a majority-American board, and — most remarkably — create a “golden share” granting the U.S. government (the filed paperwork actually names Donald Trump specifically) a board seat and veto power over corporate decisions the president deems relevant to national security. By September, Trump had already exercised it to block a plant closure in Illinois.

The legal architecture was creative. Through the Committee on Foreign Investment in the United States (CFIUS), Trump used authority that Congress had actually granted, setting conditions on a voluntary transaction rather than seizing property. Under Justice Jackson’s framework, he was arguably operating within congressionally granted presidential power.

But even the “golden share” model rested on the same assumption as Truman’s intervention: that with the right legal authority, the state could keep the furnaces burning. Truman’s 1952 order was sparked by a looming strike, but he wasn’t seizing the mills to suppress the workforce. In fact, the steelworkers largely supported the move, seeing the federal government as a more favorable arbiter than the steel executives.

In that era, the struggle was ultimately over the governance of the machinery, not the coercion of the men. Because the workers were willing to keep the mills running under federal stewardship, Truman’s gamble was a matter of law. At the end of the day, when it came to steel, the fundamental thing being fought over was control of the plants, not the workforce.

Either way, the Anthropic-Defense Department showdown will end very differently

In the summer of 2025, the Pentagon signed a contract worth up to $200 million with Anthropic, maker of Claude — the first frontier AI model authorized on classified U.S. networks. By early 2026, the partnership was in crisis. The Pentagon demanded that four of its AI providers — Anthropic, OpenAI, Google, and xAI — allow their tools to be used for “all lawful purposes,” including weapons development, intelligence collection, and battlefield operations. Three showed flexibility. Anthropic drew two bright lines: no mass surveillance of Americans, no fully autonomous weapons.

The confrontation came to a head this week. Defense Secretary Pete Hegseth summoned Anthropic CEO Dario Amodei to the Pentagon for what officials described, without diplomatic softening, as a “shit-or-get-off-the-pot meeting.”

Hegseth reportedly issued his ultimatum. Amodei arrived and, by all accounts, was cordial and unmoved in roughly equal measure. He thanked Hegseth for his service. He reiterated Anthropic’s red lines. He left.

Within hours, Hegseth’s ultimatum leaked. Anthropic has until Friday evening to grant the military “unfettered access” to Claude. If it refuses, the Pentagon will pursue one of two paths: declare Anthropic a “supply chain risk” — effectively blacklisting it from the entire Pentagon contracting ecosystem — or invoke the Defense Production Act (DPA) to compel the company to give the government unfettered access to its products on the government’s terms.

“The only reason we’re still talking to these people is we need them and we need them now,” a DOD official told reporters. “The problem for these guys is they are that good.”

That last sentence is the most important one in this story.

But let’s start with the Pentagon’s threat, which is internally inconsistent. Designating Anthropic a “supply chain risk” suggests that the company or its product are too unsafe to be within the federal government’s procurement chain. Whereas invoking the DPA suggests Anthropic’s product is so essential that the government must be able to use it. Needless to say, that’ll be a difficult split-perspective to sustain in court.

But that doesn’t mean the Pentagon can’t use these tools to cause Anthropic serious harm.

The national security apparatus can designate companies as posing unacceptable risks to U.S. information and communications technology supply chains. Such a designation would not merely void Anthropic’s Pentagon contract; it would require any company with Pentagon contracts to certify that Claude is not used in their government contract work. Amazon, Google, and Palantir, among hundreds of other Pentagon contractors, would face the choice of dropping Claude or losing their own government business.

The specific legal challenges would be novel, and Anthropic would have strong arguments that it is arbitrary and capricious under the Administrative Procedure Act: The law was designed to address security threats, not to punish domestic companies for safety policies the government finds inconvenient. But national security deference is powerful, and courts might well allow the designation to stand pending full litigation, meaning the practical damage could be severe and durable regardless of the ultimate legal outcome. The government doesn’t need a clean legal win to make this hurt.

At the same time, the Pentagon could well end up losing this game of chicken. Not only could it lose access to Claude, but other tech companies could decide maintaining Claude in their own workflows is more important than their access to government business.

Either way, expect a prolonged, high-stakes standoff.

The DPA, on the other hand, authorizes the president to require companies to accept government contracts, prioritize government orders, and allocate materials. It has been used to ensure that manufacturers produce ventilators, semiconductors, and military equipment. What it has never been used to do is compel a company’s workers to build and maintain products against their own judgment or even their conscience.1

The statute was written for mines, farms, mills, and factories, not laboratories. “Accepting a contract” means something intelligible when applied to a widget manufacturer; it is considerably more ambiguous when applied to a model that requires thousands of researchers making millions of incremental decisions to build and maintain. Even if the government could compel a license to the current version of Claude, a model frozen in time may be worth little given how rapidly the technology evolves. What the Pentagon needs is ongoing cooperation, which the law does not obviously authorize compelling.

There is also a hard statutory wall that goes to the heart of what the government actually needs: The DPA expressly exempts employment contracts from its compulsory authority. Congress, when it wrote the law, drew an explicit line between commandeering production and conscripting labor. The government can perhaps compel a license to Claude’s current weights; it cannot, under the DPA’s own terms, compel the engineers to keep building the next version. That is not an oversight. It reflects a constitutional judgment, embedded in the statute, about the limits of compelled labor — a judgment the Thirteenth Amendment places beyond Congress’s power to override in the first place. Any legal theory that tries to reach the engineers directly through some other authority runs into that amendment head-on.

Ordering Anthropic to maintain and update a product for uses it objects to also raises First Amendment concerns; it looks less like commandeering a factory and more like compelling a publisher to rewrite its editorial standards under threat of nationalization. The government can seize a building. It cannot conscript a mind. An effort to essentially do so will likely face serious questions from courts.

Artificial intelligence is about people, not companies

Which underscores the fundamental difference between AI companies and steel companies: that the obstacles to government control are more than just legal.

For the government to succeed, it doesn’t just need the courts to agree, but it will ultimately need Anthropic’s leading scientists and engineers — or if it tries to substitute in Google or OpenAI or xAI, their scientists and engineers — to design and maintain a product they may ethically object to supporting.

And here’s what Silicon Valley already knows, even if Washington hasn’t caught up: Especially in the field of AI where the leading scientists and engineers themselves are among the most concerned about the product’s ultimate capabilities, those people will walk or potentially find other ways to frustrate the government’s goals.

We’ve seen it happen in vivid, documented detail — and the story starts with Anthropic itself. Dario Amodei and several of his colleagues founded Anthropic in 2021 after leaving OpenAI over concerns about the company’s approach to safety and the pace of AI development. That departure wasn’t an anomaly. It was the first chapter of a pattern that has only accelerated since. When OpenAI’s direction drifted further from its founding safety principles, the departures didn’t trickle — they cascaded. Ilya Sutskever, OpenAI’s co-founder and chief scientist, left in May 2024. Jan Leike, who co-led the Superalignment team, followed days later, writing publicly that safety had “taken a backseat to shiny products.” By August 2024, Fortune reported that nearly half of OpenAI’s AGI safety researchers had quietly resigned. Researcher Daniel Kokotajlo walked away from $1.7 million in equity — the vast majority of his family’s net worth — because he’d lost confidence the company would behave responsibly. Most recently, just weeks ago, OpenAI researcher Zoë Hitzig quit in a New York Times op-ed over the company’s move toward advertising, warning that ChatGPT’s archive of intimate user conversations created manipulation risks that “we don’t have the tools to understand, let alone prevent.”

And it isn’t only OpenAI. Mrinank Sharma, the head of Anthropic’s own Safeguards Research team, resigned the same week with a public letter warning that “the world is in peril” — a stark signal that even inside Anthropic, the researchers most responsible for its safety culture are watching closely and willing to leave when they feel values are being compromised.

These aren’t isolated incidents. They are a live, recurring pattern showing that the researchers who build these systems have values and they act on them. If Anthropic caves to the Pentagon’s pressure, it risks triggering the same dynamic from the inside out. A $200 million contract for a company valued in the hundreds of billions is recoverable. Losing the people who made Claude the current leader in the field is not.

There’s also a concern that goes beyond engineers simply refusing, and that’s engineers who might decide they have an ethical duty to prevent the kinds of misuses Anthropic is trying to hold the line to prevent. You might call this the Galen Erso problem, named after the fictional scientist in the Star Wars movie Rogue One who, forced by the Empire to build a weapon, embeds in it a fatal flaw unknown to his employers. It’s a safe bet many Anthropic engineers are familiar with Galen Erso, a sort of cult hero in the tech sci fi fan circles from which many come. Like Erso’s situation, if the government essentially bullies frontier labs into enabling unethical things, the risk that those engineers respond by trying to sabotage or just frustrate the product in some way — and doing so in service of what they see as a higher ethical mandate — is nonzero.

The national security establishment has all manner of protocols to prevent situations in which someone might feel pressured or incentivized to act in a manner inconsistent with the government’s interests. By doing what the Pentagon is presently doing, they are arguably creating just that situation. And that says something profound about where the real power presently lies.

The Anthropic-Defense Department fight could reorient our understanding of where power lives

For most of the twentieth century, the hierarchy between government and strategic industry was not in doubt. Governments could tax, regulate, nationalize, and seize. The assets that mattered — steel, oil, uranium — existed in physical space, subject to physical control. AI sits somewhere between those old categories and something genuinely new. The government has real and formidable tools: the DPA, supply chain designation authority, export controls, and the ability to shape the market through the sheer scale of federal procurement. None of those tools are illusory, and it would be a mistake to conclude that because AI is cognitively complex, the government is necessarily powerless.

But the government’s tools are better suited to controlling access to AI than to controlling what AI does — and the distinction matters enormously.

Cutting Anthropic off from federal contracts would hurt the company and might, over time, accelerate the adoption of more compliant alternatives. It would not give the government full control over Claude. Invoking the DPA might compel cooperation but would do so through legal mechanisms whose scope is untested, against a company that would litigate vigorously and could draw on the most sophisticated legal talent in the country. And whatever mechanism the government eventually chose, it would ultimately find itself not in a standoff with a corporation but with a specific group of highly skilled scientists who have already demonstrated, repeatedly and at personal cost, that they will choose their values over their paychecks.

Hegseth walked into the Pentagon meeting with a deadline and two threats. Amodei walked out having conceded nothing. The deadline is tomorrow. The real deadline — the moment at which AI becomes so central to military operations that the government’s dependence is total and its leverage is gone — may already be in the past.

The blast furnaces never got a vote. The scientists do. Anthropic and its peers understand that. It’s not clear the Pentagon does, but it’s soon likely to find out.

Last night, in the same place where five years ago a violent mob and 147 members of Congress attempted to overturn a free and fair election, the president laid the groundwork for the Next Big Lie.

President Trump’s State of the Union address was full of falsehoods on nearly every issue (you can find some version of a fact-check by most major outlets today), but buried nearly an hour into the record-length speech was a litany of untruths about American elections. Sowing doubts about the election process is the first step in the president’s strategy to interfere with this year’s midterm elections to ensure his party maintains power, regardless of how the American people vote.

Subscribe now

Here are the five major lies the president told about our elections, debunked.

One: ‘Cheating’ is rampant in our elections

This lie about American elections might be the single most consistent position that the president holds. At least since 2012, Donald Trump has called elections that he and/or his preferred candidates lost “rigged,” “stolen,” and a “sham.” Sometimes a conspiracy between the FBI and social media companies is the primary culprit. Other times, mail-in ballots draw his greatest ire. 2020 was a bit of an “all of the above” or “throw spaghetti at the wall” style strategy. In 2026, his primary fixation — the Next Big Lie — is widespread non-citizen voting.

Just to be clear, he’s been wrong every time.

According to an election fraud database maintained by the conservative Heritage Foundation (the authors of Project 2025), from 1982 to 2024, there were only 1,546 “proven instances of voter fraud.” In 2020 alone, dozens of courts reviewed allegations of fraud and dismissed them for lack of evidence.

Multiple bipartisan audits and hand recounts in multiple battleground states confirmed the accuracy of the results.

Undermining public trust in elections with blatant lies like this lays the groundwork for the president to attempt to use the power of the federal government to change the rules and tilt the electoral playing field in his favor to guarantee his preferred outcome in November.

Read more: Elections are the essential remaining pillar of our democracy.

Two: ‘Illegal aliens’ vote in our elections

Sixty-eight. That’s the number of noncitizens who voted between 1982 and 2024, according to the American Immigration Council’s review of the Heritage Foundation data. 68 out of over one billion total votes cast during that time. That’s below 0.0001%.

According to a Brennan Center report, in the 42 jurisdictions they studied, the total number of undocumented immigrants who were referred for investigation for illegal voting in 2016 was just 30 — 0.0001% of votes cast that year.

There’s just no evidence to support the president’s claims. But that’s beside the point for Trump. He’s searching for a pretext to interfere in election administration, stifle political activity, and silence voters so that he can control who holds office, regardless of who wins the election.

Read more: Free speech, ICE, and free elections in 2026.

Three: Mail-in voting is ‘crooked’

Ever a favorite target of the president, mail-in voting remains a convenient, secure way for Americans to vote. Take it from a recent Brookings Institution report:

We find that mail voting—universal vote-by-mail in particular—has substantial benefits for election administration and tens of millions of U.S. voters. As part of this analysis, we corroborate prior research that mail voting is secure and find low overall mail voting fraud, particularly in universal vote-by-mail voting systems.

To put a finer point on it, cases of mail-in voting fraud account for only 0.000043% of all mail ballots cast. That’s four in 10 million.

Most Americans want every eligible voter to be able to vote and have their vote counted. Most of us support absentee and early voting, and many of the president’s own voters rely on these options to cast their ballots.

Despite President Trump’s vague reassurances last night, restricting absentee voting primarily impacts military voters, seniors, voters with disabilities, and rural communities who rely on mail voting to make their voices heard. And all of us benefit from early voting options. Eliminating ballot drop boxes and limiting in-person early voting would create longer lines and more chaos on Election Day, which could turn away even more voters.

Four: The SAVE America Act is ‘common sense, country-saving’ legislation

This sweeping “election integrity” bill is an effort by an increasingly unpopular president and party to pick and choose who gets to vote in a desperate attempt to stay in power.

Instead of preventing rampant non-citizen voting (which, as I discussed above, isn’t real), the impact of this legislation would be to make it harder for law-abiding citizens to vote, requiring all of us to find or pay for expensive documents, made even more complicated for anyone who has changed his or her name — including the vast majority of married women — to register to vote. And the registration process itself may become prohibitive for many voters; the bill could effectively end or severely limit automatic, mail-in, and online registration methods.

Americans deserve simple, accessible systems that allow every eligible voter to participate in our democratic process and have their vote counted.

Right now, roughly 21 million eligible citizens do not have immediate access to the specific documents this bill requires. For many Americans, obtaining these documents requires time, travel, and money. And it’s important to note that this legislation will disproportionately impact Black, Latino, Asian, Indigenous, and immigrant communities.

When viewed alongside this administration’s ever-escalating attacks on free and fair elections, it becomes clear that the SAVE America Act is part of a coordinated effort to silence voters who the president expects to vote against his party.

Moreover, because the bill is unlikely to pass over a Democratic filibuster, its failure may also be used as a pretext to attack future election results. An equally dangerous scenario for the SAVE America Act isn’t just that it could pass and disenfranchise millions of voters. It’s also that it could fail — and have that be used to try to invalidate everyone’s votes.

Read more: Trump’s elections Chimera.

Five: Voting is a privilege

Of all of the falsehoods uttered by the president last night, this one may have upset me the most (lots of competition for this, I know), even if it largely flew under the radar for most commentators.

Voting isn’t a privilege; it’s a constitutionally protected right of American citizens. Full stop.

When President Trump called voting a privilege, he said the quiet part out loud. If the government views voting as a privilege that it can pick and choose upon whom to bestow, no American’s right to vote is safe — even those who don’t expect to be impacted by the president’s current disenfranchisement efforts.

This year, on the 250th anniversary of our country’s founding based on the principle that citizens have a God-given right to choose their own leaders, we have the opportunity to reflect on the struggle and sacrifice that have expanded the right to vote. The Women’s Suffrage and Civil Rights movements. The bloodshed on battlefields and on American streets. The leadership of activists, soldiers, teachers, union organizers, brave citizens from all walks of life. And in this reflection, we can find inspiration to keep going as we take up this mantle.

This year, we are engaged in a struggle to defend that most fundamental right of Americans to govern ourselves. Last night made clear that the challenge we face is steep, but I have no doubt that we will win.

The U.S. Department of Education’s recent decision to investigate the National Study of Learning, Voting, and Engagement (NSLVE) didn’t make many national headlines, but perhaps it should have.

Subscribe now

For over a decade, NSLVE has been the gold standard for nonpartisan research into college civic learning and engagement. This investigation — thinly veiled as a concern for student privacy — threatens to dismantle an essential academic framework. By centering on alleged violations of the Family Educational Rights and Privacy Act (FERPA) — the federal privacy law carrying the same institutional peril as a HIPAA violation in medicine — the department is launching a calculated assault on the right of universities to measure their own civic and educational outcomes.

Is measurement a crime?

American universities are required by law (the Higher Education Act of 1965, as amended) to help students register to vote. But the effort to engage students in democracy on campus is driven by more than just a legal mandate. Many institutions explicitly name civic knowledge and democratic participation as central to their educational mission. Just as a chemistry department uses exams to measure a student’s subject matter expertise, a university must use data to measure a student’s mastery of participation in our democracy. Before NSLVE, which was launched in 2013, colleges and universities had no empirical way to measure the success of their civic engagement efforts. NSLVE has bridged this gap and provides data-driven insights to over 1,000 U.S. colleges and universities, allowing them to understand and improve student learning and civic participation on their campuses.

This study has transcended partisan lines, earning the endorsement of both Democratic and Republican administrations alike. For years, the federal government has actively encouraged participation in this study because NSLVE has proven to be a trusted partner with a process explicitly designed to be FERPA compliant.

NSLVE’s structure and policies have always prioritized student privacy. In fact, NSLVE never handles identifiable student information. It utilizes de-identified records from the National Student Clearinghouse and has no way of knowing if or how any specific student voted. The study does not share individual student data with any participating campuses or political or non-political third parties. And results are strictly limited to national trends or confidential campus-level reports, ensuring that no individual student can ever be identified.

A pattern of anticipatory obedience

Given this structure, the department’s claim that NSLVE data is being used by “political organizations to influence elections” is functionally impossible. But that impossibility is beside the point. By threatening to pull federal funding, the administration is seeking to create a state of anticipatory obedience, a move pulled straight from the authoritarian playbook.

“Creating an environment of fear and division is key,” writes our colleague Shanna Singh Hughey in her piece on anticipatory obedience. “People under threat naturally do what they can to avoid injury and maximize their chances of survival. And they do the same for the organizations they run.”

Read more: Reversing the vicious cycle of anticipatory obedience.

In this case, when the Department of Education warns that any school using this year’s data “could be at risk” of a FERPA violation, they aren’t seeking a legal victory; they are seeking a chilling effect. They want university lawyers to preemptively pull out of the study rather than risk their entire budgets on legal battles or cuts in federal funding. If they succeed, it will force a massive retreat from civic programs in the middle of a major election year, effectively gutting the infrastructure of youth participation in our democracy.

This is not an isolated incident. It is the latest in a series of moves designed to make it harder for universities to promote nonpartisan civic engagement. Last summer, the Department of Education prohibited universities from using federal work-study funds for nonpartisan voter engagement. And on the very first day of this administration, the president revoked the executive order that had directed federal agencies — including the Department of Education — to promote nonpartisan voter registration. It’s hard to view these actions as anything other than the administration trying to prevent more voters from voting.

The impact is the intent

Of course, if the real goal is to create a climate of fear, the legitimacy of this investigation would be irrelevant to its authors, who are seeking to delegitimize civic life and reframe the simple act of voting as a partisan threat. Whether through the quiet bureaucracy of a FERPA probe or by threatening the presence of ICE officers at the polls, the intent remains the same: to erect as many barriers to participating in our democracy as possible.

So while this technical investigation into data privacy may not have dominated national headlines, we still need to pay attention. This serves as a critical data point in the administration’s escalating use of the vast powers of government to target its perceived political opponents and ongoing efforts to suppress participation, and a reminder that not all authoritarian election rigging happens out loud. The current administration is using every small lever, every instrument at their disposal to choose its voters, out of fear that voters would choose their opposition.

ICE wants protesters to know they’re being watched — and it has the technology to back it up.

In Minneapolis, ICE officers scanned protesters’ faces at demonstrations, then warned that the protesters were going to be added to a government database. In St. Paul, an agent photographed a vehicle’s license plate and used that information for intimidation — not just by addressing the legal observer by name, but by driving onto the street where she lives. In Maine, ICE officers showed up at an observer’s home and warned, “We know you live right here,” after an encounter at an elementary school bus stop.

Their message is clear: We know who you are. We know where you live. We’re watching.

Subscribe now

These aren’t isolated incidents. And while they are shocking, they tragically aren’t surprising. These examples are the byproduct of a federal law enforcement agency with dramatically expanded access to surveillance technology acting on behalf of a Trump administration that has decided constitutional rights are threats to be neutralized.

The playbook is familiar to anyone who has studied authoritarian regimes: Identify dissenters, make them visible to the state, and demonstrate that opposition has consequences. To be clear, the United States has a sordid history of surveilling activists, a history this administration is undeniably building upon. DHS Secretary Kristi Noem has equated those protesting and observing ICE with terrorists, calling the filming of ICE operations a form of anti-ICE violence. Immigration officers are reportedly being directed to gather identifying information on anyone who films them.

But it’s not just the administration’s equating protesters to domestic threats that has led to the scenes we’re witnessing in Minnesota and Maine — it’s ICE’s expanding surveillance arsenal. Under the Trump administration, ICE has become the most well-funded law enforcement agency in the country, and it’s putting that funding to work to build a data collection apparatus to match. While ICE’s surveillance capabilities have grown under administrations led by Democrats and Republicans alike, what sets the second Trump presidency apart is its aggressive deployment of technology — supercharged by AI — to suppress constitutionally protected activities. ICE officers in the field currently have tools to identify protesters, map their associations, and find out where they live — all in real time.

As protesters across the country oppose ICE’s immigration crackdown and documented violence against immigrants and citizens alike, the agency is increasingly demonstrating a willingness to turn these tools against Americans exercising their First Amendment rights.

The function isn’t security. It’s suppression.

But we’re not powerless. ICE’s surveillance apparatus is in part reliant on access to state databases and local partnerships. Around the country, state and local communities are beginning to cut off some of those access points that are ripe for abuse (more on this below).

ICE’s surveillance arsenal

ICE’s toolkit for identifying and intimidating protesters includes a vast array of technologies, many of which remain shrouded in secrecy. But what we do know about how they’re being deployed against protesters is alarming.

Facial recognition technology: ICE officers in Minneapolis reportedly scanned protesters’ faces at the scene where Alex Pretti was killed by federal agents. Observers across the Twin Cities have reported that ICE agents are warning that they’re using “facial recognition” technology and photographing the faces of residents.

Contract records and internal ICE emails reveal the agency is using at least two facial recognition tools. The first is Mobile Fortify, an AI-enabled smartphone app ICE officers are using in the field to verify citizenship status and scan protesters’ faces. The second is Clearview AI, which signed a new contract with ICE in September to provide access to its massive database of facial images scraped from the internet — in part to investigate “assaults against law enforcement officers.”

Automated license plate readers (ALPRs): ICE agents eerily driving to the very St. Paul street where a legal observer lives may have been enabled by the weaponization of ALPRs. Companies like Flock Safety and Motorola Solutions operate cameras across the country that can be used to track vehicles’ location histories and identify their owners. Until recently, federal agencies, including ICE, had access to Flock’s national lookup tool, which aggregates data from thousands of local police departments using the company’s AI-enabled ALPRs. ICE has also begun using a smartphone app called Mobile Companion to scan license plates that’s integrated with Motorola Solutions’ ALPR network. The result: ICE can determine who protesters are, where they live, where they’ve been, and who they associate with.

ICE claims broad authority to deploy these tools, with DHS defining “violence” to encompass making and posting videos of agents in public — activities that ICE might then investigate as “assaults” despite generally being protected by the First Amendment. In a September presidential memorandum, the administration further named “anti-Americanism” and “extremism on migration, race, and gender” as potential motivations animating “violent conduct.” Essentially, any American who observes, protests, or records ICE operations can be subject to AI-powered surveillance.

Government data is expanding a panopticon

These surveillance tools all derive their power from data — including sensitive personal data that Americans have already entrusted to their government for entirely different purposes.

ICE has direct access to a range of data that states collect on their residents through the International Justice and Public Safety Network, more commonly known as Nlets, a data-sharing network for law enforcement agencies across the U.S. When ICE runs license plates through the Nlets system, for example, it can quickly gather sensitive personal information, including residential address, date of birth, physical characteristics, license plate, vehicle registration information, and facial photos (especially helpful for utilizing facial recognition technology).

As reported by 404 Media, Mobile Fortify integrates data from Nlets and several federal databases. And through the ALPRs described above, ICE can access data that is collected by local and state police for routine law enforcement.

This interconnected web of data creates a surveillance ecosystem where a single data point — a license plate, a face in a crowd — can unlock a comprehensive profile of an individual’s identity, location, and associations. And the reality is that government data enables much of (though not exclusively) this capability.

That is why, as we wrote last May, the Trump administration has been on a quest to centralize all forms of data collected by the government. ICE’s overreaching use of state and local law enforcement data fits the same pattern of systematically leveraging data collected for legitimate government purposes — tax compliance, benefit programs, routine policing — and repurposing it for surveillance and crackdowns without public notice or meaningful oversight.

As the agent of this administration, ICE wants us to believe that surveillance is inescapable. However, the tools that enable federal agents to identify and intimidate Americans engaged in First Amendment-protected activity depend in part on state data and local contracts. When communities put bottlenecks on the data supply, ICE’s surveillance apparatus weakens. ICE’s power isn’t absolute. It’s borrowed — and we can take our data back.

What you can do to stop it

Just as data provides the fuel for ICE’s advanced surveillance technology, so too does it provide leverage for opposition. From deep blue cities to red state legislatures, public pushback against ICE’s surveillance tactics is growing, and it’s working. The key is to reclaim power over our data.

Local and state governments are taking measures to curtail ICE’s weaponization of their constituents’ data to target and intimidate citizens and immigrants alike. Over 15 cities across at least 12 states have terminated or otherwise ended their contracts with surveillance giants like Flock Safety, which appears to have misled its customers about DHS’s access to customer data via their “national lookup” tool. Five states have severed ICE’s access to DMV data through Nlets.

Of course, none of these measures alone is a silver bullet. ICE undoubtedly will seek workarounds and loopholes to gain access to some of this data for unconstitutional purposes. But the reason these measures are powerful is that they provide openings to limit the surveillance machine. While they won’t deter a federal government bent on violating the First Amendment, they do slow and inhibit the weaponization of data against law-abiding citizens, buying time and space to engage in dissent.

Here are three ways you can help curtail ICE’s abusive applications of data and technology:

1. Keep ICE away from DMV data.

Many states provide ICE with unrestricted access to their DMV data through Nlets, allowing ICE agents to gather troves of information on individuals in the field — without a warrant — using only license plates. But some have already moved to change this.

Illinois, New York, Massachusetts, Minnesota, and Washington have already blocked ICE from accessing their DMV data via Nlets. New York has a clear law prohibiting motor vehicle offices from sharing information with ICE for civil immigration enforcement, while Massachusetts’ nondisclosure regulation came from its attorney general’s rulemaking. Minnesota has reportedly stopped sharing all motor vehicle data with ICE (even though its driver privacy law more narrowly restricts certain data sharing, specifically non-REAL ID data), and Oregon is reportedly taking a similar step to ensure its residents’ data is not available to ICE through Nlets without a legislative requirement to do so.

Encourage your state to join their ranks. Contact your state secretary of state’s office, governor’s office, and elected representatives to encourage them to take action to block ICE’s access to your DMV data through Nlets.

Use this email template to send a message in under two minutes.

2. Interrogate local connections to companies like Flock Safety and Motorola Solutions that have a history of sharing ALPR data with ICE.

Local pressure is driving cities to re-evaluate their data-sharing agreements or even terminate their contracts with ALPR companies like Flock Safety. In Flagstaff, Arizona, the City Council unanimously voted to cancel the city’s contract with Flock Safety following extensive outcry and activism from the community. And just last week, 700 people gathered on the plaza in front of Bloomington, Indiana’s city hall to protest the city’s contract with Flock Safety.

Equip your city council members or local sheriff with information on how ALPR data is being used by ICE. Show up to a town meeting, or consider requesting a one-on-one meeting to discuss your town’s ALPR contract.

Download this ALPR advocacy toolkit from the ACLU of Iowa to start a conversation with your city council.

3. Make the invisible visible.

Surveillance thrives in the shadows — and nothing busts shadows quite like a little sunlight. Shedding light on ICE’s surveillance technology and the data it runs on is a vital first step toward building the community power necessary to demand change.

Contact your local newspaper to pitch a story or letter to the editor to expose how your town’s tax dollars may be funding federal surveillance.

Check out the Atlas of Surveillance (a project by the Electronic Frontier Foundation) to see which ALPR technologies are currently deployed in your town. Use this resource to search for your town and select “automated license plate readers” to see results about which ALPR technologies are present in your community. Pair that information with the ACLU of PA’s Guide to writing an LTE to start the conversation.

In the timely Star Wars spinoff Andor, the young rebel Karis Nemik writes, “Tyranny requires constant effort. It breaks, it leaks. Authority is brittle.”

This is the lesson of history. Authoritarianism is not new. It has appeared throughout human civilization. But we live in a largely free world today because of Nemik’s observation: Authoritarianism has been — at least historically — unstable. It has relied on imperfect information, fallible human enforcers, and coercive systems that strain under their own weight. And because of that brittleness, people have repeatedly been able to correct away from tyranny and toward freer, more democratic societies.

But what if that were to change?

Subscribe now

In his recent essay, “The Adolescence of Technology,” leading AI lab Anthropic’s co-founder Dario Amodei opens with a different science-fiction reference: a scene from Carl Sagan’s Contact meant to illustrate a sobering idea — that societies capable of achieving immense technological power must also learn how to survive it.

Among the usual concerns of AI risk assessors that Amodei includes — human extinction, human subjugation to misaligned AI, mass unemployment — is a concern that needs far greater amplification and attention: the possibility that AI could undermine the historical dynamic Nemik describes. The very dynamic that has been democracy’s salvation. That AI could harden authority rather than expose it. That it could seal the cracks through which freedom has always eventually re-emerged.

In short, Amodei is warning that powerful artificial intelligence may be the first technology in human history capable of making authoritarianism permanent — not by violently overthrowing democracy but by eliminating the possibility of democratic reversal.

If that is correct (and, as Amodei acknowledges, there are still plenty of uncertainties that remain regarding both whether AI will achieve such advanced capabilities and the speed at which this occurs), we are not merely facing another technology policy challenge. We are approaching a civilizational threshold.

What’s more, while Amodei’s assessment largely looks to a future moment when AI has become powerful enough to be the equivalent of a “country of geniuses,” the harms caused by AI in the hands of authoritarians are all too real today. If we are to preserve the possibility of democratic reversal, we must act now.

Why authoritarianism has always failed

Authoritarian regimes have always appeared formidable from the outside. Yet their defining feature has been fragility.

They suffer from information problems. Fear distorts reporting; loyalty replaces truth. The “Dictator’s Trap” is that everyone around an authoritarian is afraid to deliver bad news and so provides him with misleading information that leads to overreach and error.

Authoritarians rely on human agents who hesitate, defect, or leak and who have capacity constraints by dint of their very humanity. They must tolerate inefficiencies and often breed and encourage corruption to maintain control.

And because repression is costly — politically, economically, and psychologically — it tends to provoke resistance, fracture elites, and invite external pressure.

Even the most brutal regimes of the twentieth century eventually broke. Fascist states collapsed under war. Communist regimes stagnated, splintered, or reformed. Military juntas gave way, sometimes suddenly, sometimes haltingly, to civilian rule.

The lesson is not that democracy is inevitable. It is that authoritarianism has never been able to fully close the loop. Something always leaked.

Amodei’s warning is that “powerful AI” could change this equation:

Current autocracies are limited in how repressive they can be by the need to have humans carry out their orders, and humans often have limits in how inhumane they are willing to be. But AI-enabled autocracies would not have such limits.

Authoritarian lock-in

Amodei identifies a convergence of sufficiently powerful AI-enabled capabilities that, taken together, threaten to eliminate the mechanisms through which authoritarian systems have historically failed:

• Total surveillance: Systems become capable of ingesting vast amounts of communication, movement, and behavioral data so that there’s no escaping or hiding from the controlling apparatus of the state. Resistance can be identified and rooted out before it can bear fruit.

• Personalized propaganda: Persuasion is optimized at the level of the individual, leveraging data on emotional states, social context, and psychological vulnerabilities collected by scraping data from digital (and potentially even real world) interactions. Individuals can be manipulated into consenting to giving up their freedom, rights, and willingness to dissent.

• Autonomous weapons: The state has the ability to enforce suppression at scale through automated tools capable of responding with full force to every transgression and every minor act of dissent or organizing — and to do so potentially autonomously and thus unconstrained by the capacity and accountability limits of human bureaucracies.

• Strategic mastery: With a “country of geniuses,” a leader or entity bent on power could strategically outmaneuver any not similarly equipped opposition, domestically or internationally, extending indefinitely the totality of their control.

Together, these capabilities point toward what can be described as authoritarian lock-in: a condition in which a regime does not merely suppress opposition but structurally prevents opposition or displacement from ever becoming effective.

The danger is that powerful AI could seal authority rather than strain it — closing the feedback loops through which societies have historically corrected themselves. Once these feedback loops are closed, the “adolescent” society never matures; it simply ossifies.

This concern differs in a significant way from concerns about human extinction or human subjugation at the hands of powerful AI. Whereas those concerns, however serious (and they are), are speculative and await the advancement of powerful AI, the concern about AI being used to empower authoritarian suppression is already here.

China is already using AI for facial recognition to track Uyghurs, for predictive policing to identify potential dissidents, for content moderation to suppress speech at scale. They are exporting this technology to other authoritarian regimes. And here in the U.S., the current administration is already implementing some of these tools.

The question that follows: What do we do about it?

Within the AI community and U.S. political community, there is already a dominant answer to this question.

Most leaders and thinkers closest to frontier AI — including Amodei himself — identify the primary danger as the possibility that China develops powerful AI first.

The reasoning is straightforward. China is already an authoritarian state. If it acquires powerful AI systems capable of total surveillance, predictive repression, and overwhelming strategic advantage, it would not merely entrench authoritarian rule at home. It could utilize its AI-enabled power advantage to extend its control abroad, potentially even achieving global hegemony under a totalitarian umbrella. Authoritarian lock-in, but without borders.

Fear of China’s global hegemony augmented by these tools often points to a single overriding imperative: The United States must beat China to the most powerful AI. The Trump White House has put it in those terms, declaring that “to remain the leading economic and military power, the United States must win the AI race.” On Capitol Hill, the message is even more explicit. Senate Commerce Chair Ted Cruz has argued that “the way to beat China in the AI race is to outrace them in innovation,” urging policymakers to “remove restraints” that slow development. And at hearings convened under the banner of “Winning the AI Race,” the framing is blunt: “America has to beat China in the AI race.”

Industry leaders echo the same logic: Microsoft President Brad Smith told lawmakers that the “number one factor” determining whether the U.S. or China “wins this race” is which technology is adopted globally, warning that “whoever gets there first will be difficult to supplant.” Former Google CEO Eric Schmidt has repeatedly warned that AI leadership will shape the global order. The result, as the Washington Post describes, is a “near-consensus” among senior officials and top executives that the U.S. must let companies “move even faster” to maintain its edge over China.

It is for this reason that Amodei himself describes advanced chip export controls as among the most important interventions available to democratic governments. The underlying theory is clear: If AI is dangerous, it is far more dangerous in the hands of an authoritarian state than a democratic one.

This assessment has profoundly shaped the orientation of the U.S. AI ecosystem.

The perverse dynamic it creates

The China-first framing is not wrong. But it is dangerously incomplete.

If beating China to powerful AI is the overriding objective, the logical conclusion is speed at all costs. Minimize regulation, remove friction, treat guardrails as liabilities, and view caution as complacency. This logic pushes U.S. firms to race ahead, consolidate power, and resist democratic constraints — all in the name of preventing authoritarian lock-in abroad.

But that same logic leaves us exposed to a second, equally serious risk: authoritarian lock-in at home. Amodei himself recognizes this risk, though not specific to the U.S., in acknowledging that “a hard line” must be drawn “against AI abuses within democracies” with “limits to what we allow our governments to do with AI, so that they don’t seize power or repress their own people.”

Last year, one of us spoke with a founder of a leading AI lab. When we asked what worried them most, their answer was immediate: not China, but the possibility that extraordinarily powerful AI systems might come online while a leader with a deeply instrumental view of power controls the U.S. government.

This was not a partisan remark so much as a structural one. AI dramatically lowers the cost of surveillance, enforcement, and control. It centralizes authority. It weakens accountability. Those effects are dangerous in any system — but maximally so when wielded by leaders who already reject institutional constraints.

And as President Trump said just this week, “I don’t need guardrails. I don’t want guardrails. Guardrails would hurt us.”

The domestic authoritarian risk

The Trump administration largely accepts the China-first theory. It has embraced AI as a strategic asset and treated governance as an impediment to innovation. It has sought to remove guardrails in the name of competitiveness and national strength.

At the same time, our government is led by a president who has repeatedly demonstrated contempt for checks on his own power — attacking courts, undermining independent oversight, threatening political opponents, and praising strongman tactics. The administration has pursued expanded surveillance authorities, suppression of dissent, and consolidation of executive power — directionally aligned with the very risks Amodei warns about. Ironically given a strategy ostensibly grounded in concern about what China would do with powerful AI, Trump has praised Xi Jinping’s leadership style, calling him a “brilliant” leader for “control[ling] 1.4 billion people with an iron fist.”

The uncomfortable reality is this: The problem is not only which nation gets powerful AI first but who controls that nation when it arrives.

If AI hardens authority, then leaders tempted by authoritarian methods, whether in Beijing or Washington, pose a similar structural threat. If we have one quibble with Amodei’s essay, it’s that it seems to underplay this domestic risk, perhaps because naming it explicitly carries costs for an American company right now. But that itself underscores the problem. And so we’re taking one of Amodei’s pieces of advice (“the first step is … to simply tell the truth”) and naming that risk more explicitly here.

Political economy makes this worse

Layered on top of this dilemma is a brutal political economy problem.

Frontier AI development is extraordinarily capital intensive. Compute, data, and talent are concentrated in a handful of firms. Those firms command immense economic power — and increasingly, political influence.

As Amodei notes, AI is such a powerful economic and geopolitical prize that the risks of regulatory capture are intensified. Already, more than $100 million has been committed to an industry-backed super PAC committed to preventing regulation, and the pool of capital available to augment that is virtually bottomless. Democratic checks, already strained, risk being overwhelmed not just by urgency and fear of falling behind but by a huge campaign finance warchest.

The result is a convergence of concentrated private power and concentrated state power at precisely the moment democratic safeguards matter most.

Amodei acknowledges this problem but doesn’t offer a realistic proposal for solving it. If there’s a glaring failure in his essay, that is it.

The only durable solution

If AI can produce authoritarian lock-in, then preventing it requires more than choosing the “right” geopolitical winner or slowing development “slightly” (to use Amodei’s term) at the margins. It requires actively designing AI and its governance to advance democratic practices and values.

The only durable alternative is not merely to anticipate how AI may harm democracy and impose penalties for its abuses and harms. Instead, we must account for AI’s risks to democracy in its development and deployment while using it to affirmatively make democracy work better.

That means AI systems that increase transparency rather than secrecy; that strengthen accountability rather than weaken it; that distribute power rather than concentrate it; that help citizens understand, deliberate, and participate rather than manipulate or surveil them.

These goals must be built into the design choices being made regarding AI now, at the market level, the corporate level, and the governance level. They need attention, investment, and innovation – today.

From insight to institution: the AI for Democracy Action Lab

This recognition is why Protect Democracy launched the AI for Democracy Action Lab.

What is missing in the AI ecosystem is not analysis, white papers, or technical benchmarks, nor efforts focused on existential risk to humanity (though those are important too). What’s missing are institutions explicitly dedicated to defending democratic reversibility — the capacity of societies to correct course when power is abused. And getting there means building AI in a way that both anticipates and wards off the risks of authoritarian lock-in while advancing and improving democracy and democratic commitments.

When one of us spoke to another top executive at a frontier lab recently and asked about the tenor of conversations in their c-suite and lunchroom about how AI might shape the battle between democracy and authoritarianism, their answer was “what conversations?” While that might have been hyperbolic to make a point, it illustrated an imbalance that Amodei’s essay nobly strives to correct. We must invest far more attention and energy on the risk AI poses to democracy and democratic reversal.

That’s what the Lab is designed to do by focusing on three fronts:

1. Defending democracy from AI-enabled authoritarianism through litigation, regulatory measures, and effective governance.

2. Ensuring AI contributes to a healthy civic media ecosystem.

3. Harnessing AI to strengthen democracy by creating products that augment democratic practices.

This work is not only for civil society and government. It is the real challenge for AI companies themselves. For Anthropic and others who take Amodei’s warning seriously, avoiding authoritarian lock-in cannot be a secondary concern. It must be a design principle.

The choice before us

Nemik was right: Tyranny has always required constant effort. It has broken because it is brittle.

As Amodei argues, powerful AI threatens to eliminate tyranny’s inherent vulnerability. It threatens to seal authority, automate control, and make correction impossible.

But, crucially, today we are not simply awaiting the other shoe to drop. The AI of today is already redefining authoritarian capabilities inside and outside the U.S.

Whether AI will either make democratic governance more capable than it has ever been — or render it obsolete — will not be decided by the technology itself. That question will be decided by us as AI’s shapers and wielders. We are the determinative factor in whether action is taken before authoritarian lock-in becomes irreversible.

We have argued that the contest between democracy and authoritarianism is the test of our time, and it is now clear that one of the central fronts of that contest will be which side AI advantages.

It is up to us to get that right before it’s too late. The clock is already running.