A Week's Catch-Up

With apologies for the lapse

Good Evening:

This red fox in my neighborhood sincerely apologizes for not having kept up the daily part of dog shirt daily of late. He’s had a lot going on, but he’s most sincerely trying to get back into a regular rhythm.

---

Monday on #DogShirtTV, the estimable Holly Berkley Fletcher and I opened a Greek Chorus discussion about what mischief can be done with the reflecting pool:

Tuesday on #DogShirtTV, the estimable Anna Bower showed up with a long tale of woe, and we discussed how to turn it—and her life, more generally—into a romantic comedy. She even had a jaunty rom com hat for the occasion:

Wednesday on #DogShirtTV, the estimable Alicia Wanless, the estimable Holly Berkley Fletcher, and I celebrated Canada Day—even though it was not Canada Day:

Thursday on #DogShirtTV, the estimable Anastasiia Lapatina brought us the craziest news stories from Ukraine, which included Monaco assassination attempts, murders, secret meetings, and more:

And Friday on #DogShirtTV, I brought on the estimable Ariane Tabatabai, fresh from the NATO summit in Turkey, to tell us all about it:

---

Recently On Lawfare

Compiled by the estimable Sarah Willrich

Reconstructing Iran’s Nuclear Baseline

In the latest edition of Lawfare’s Foreign Policy Essay series, Bailey Schiff argues that the gap in baseline knowledge about Iran’s nuclear capabilities has widened since Iran ended Additional Protocol monitoring in 2021 and then 2025-2026 U.S. and Israeli strikes on its nuclear sites. Schiff highlights that the new memorandum of understanding with Iran represents a critical juncture in which the U.S. can push for monitoring and verification measures that fill that gap—or risk cementing the present ambiguity for years to come.

Even with expanded access, reconstructing what happened during monitoring gaps would be a massive undertaking that would require full Iranian cooperation and a verification budget far greater than the pre-JCPOA level. It would first require an Iranian declaration, followed by IAEA access to conduct damage assessments and item-by-item counting and tagging of rotors and bellows, as well as the machines that produce them. Even then, the effort would likely remain incomplete, in part due to the loss of institutional knowledge on the Iranian side. Strikes likely degraded some records at nuclear sites, key scientists have been assassinated before and during both military operations, and Iran likely compartmentalized its nuclear program to prevent leaks, restricting access to a small cadre of specialists. As a result, even a fully cooperative government may be unable to reconstruct a complete account of its own activity.

Afrikaner Refugees and the Limits of U.S. Pressure on South Africa

Mattie C. Webb places the Trump administration’s exceptional-refugee-status policy for Afrikaners into the historical context of apartheid-era solidarity networks. She suggests that President Trump’s support of the “white genocide” myth is undermining U.S. credibility abroad and hardening an already deteriorating relationship with Pretoria, which rejects the narrative and has instead used the standoff to burnish its own standing as an independent middle power rooted in Global South solidarity.

South Africa has proved resilient in the face of U.S. criticism, and defending its domestic agenda is consistent with its wider foreign policy. South African leaders remember both the solidarity they received from the Global South during apartheid and the ambivalence they often expected from Washington, which frequently resisted the burgeoning anti-apartheid movement’s pressure to impose mandatory economic sanctions on South Africa. This acute memory makes the ANC’s positions on Palestine, Iran, and China seem consistent, even when they run counter to U.S. interests.

Accusation, Trust, and the Future of Vulnerability Disclosure

Cory Simpson examines Microsoft’s ambiguous threats of legal action against the researcher Nightmare Eclipse, who identified unpatched vulnerabilities in Windows systems. Simpson argues the episode exposes an industry-wide problem in coordinated vulnerability disclosure: eroding trust and communication between companies and the researchers who help keep their products safe.

CVD is not whistleblowing, but the two systems share a basic design principle: People are more likely to surface risk when they trust the channel and believe they can use it safely. That design principle dates back to the earliest years of the republic. In 1778, the Continental Congress protected sailors and marines who reported misconduct by the commander of the Continental Navy and later faced a criminal libel suit. The lesson still applies. Systems that depend on disclosure must protect those willing to come forward. Researchers who fear legal exposure may stay silent, delay reporting, or disclose outside the vendor process.

America Won’t Beat the Distillation Ecosystem

In the latest edition of the Seriously Risky Business cybersecurity newsletter, Tom Uren discusses Anthropic’s allegation that Alibaba conducted a major distillation attack that used Anthropic’s Claude to train its own artificial intelligence (AI) model, reports that last year’s hack of Jaguar Land Rover was conducted by Russia, the restoration of Anthropic’s Fable model, and more.

This entire ecosystem is financed by a whole range of users, not just AI companies looking to conduct distillation attacks. Users include university professors and students, tech workers, developers, resellers who buy wholesale access and repackage it for individual consumers, and even hobbyists.

Access to Claude via these means is amazingly cheap, too, with Chinese users paying as much as 70 percent to 90 percent below official prices. Transfer stations achieve a cost advantage by, among other things, harvesting bulk registered free sign-up credits and even selling user logs of requests and responses to be used by Chinese model makers for distillation.

Gradually, and Then Suddenly: The Decline and Fall of ODNI

Michael Feinberg and Julia Curlee trace the current politicization at the Office of the Director of National Intelligence (ODNI) back through the office’s larger history. They explain that ODNI’s vague authorities, created in 2004 to coordinate a fractured post-9/11 intelligence community, were historically constrained by political norms to remain largely administrative. Feinberg and Curlee argue that those same authorities now give former DNI Tulsi Gabbard and acting DNI Bill Pulte the tools to manipulate intelligence and potentially interfere with elections.

It is entirely fair to ask how much harm a DNI, with a reduced staff of roughly 1,300, can actually do. For the bulk of its life, the ODNI did not cause any harm. Given that it was founded in response to a national emergency, a hastier founding than many other departments’ creations, that is no small thing. But its impact has been, at best, modest. A review of its published directives reveals that they are largely administrative in nature, and many of them simply direct component agencies to draft their own policies in response to problems identified by the ODNI. The organization and its director functionally manage administrivia, deconflict intraagency problems, and collate information. How much damage could putting nonprofessionals in charge of such an enterprise actually cause?

Quite a lot, as it turns out.

Can Government of and by API Still Be Government ‘for the People’?

James Goodwin and Arvind Salem argue that the Department of Defense’s designation of Anthropic’s top artificial intelligence (AI) models as a “supply chain risk”—barring Pentagon contractors from working with the company—reveals an even bigger issue than the immediate legal dispute: how dependent the government, and much of the private sector, has become on just a few AI companies. That dependence, Goodwin and Salem contend, further removes decision-making power from civil servants accountable to the public.

In this sense, AI can be understood as combining two developments that have gradually eroded civil service influence: privatization and the political power of private vendors. Policymakers from both parties have long embraced contracting out as a way to shrink government and make it “run more like a business.” And the close relationship between the Defense Department’s political leadership and the network of private defense contractors has long shaped national security policy, often at the expense of career expertise. What AI represents is a second wave: Whereas the first replaced public-sector workers with private-sector ones, this wave effectively replaces both with privately developed systems.

The Military and Elections, Part I: The Legal Wall

In the first part of their analysis on the potential for deploying the military during elections, Natalie Orpett, Molly Roberts, and Loren Voss unpack the key statutes preventing such deployment, from the Civil War era through Reconstruction to modern limitations. They explain that the history is less a straight line than a fight: Congress moved to keep troops out of polling places as early as 1865, only to expand the military’s election role during Reconstruction, before reasserting and tightening the restrictions that remain on the books today—restrictions that go beyond the standard limitations on domestic deployments.

The history of elections-related domestic deployment law only reinforces the view that Congress wanted to keep the military out of the voting process. The Civil War and Reconstruction were essential to the evolution of these statutes. The period was highly anomalous: For many of those years, Congress made more room for military involvement in elections. But the way this disruption was resolved actually helps explain some of the restrictions still in place today.

The Military and Elections, Part II: Deploy First, Litigate Later

In the second part of their analysis, Orpett, Roberts, and Voss consider alternative interpretations of these statutes and the Constitution that claim the president is not barred from deploying the military during elections under certain circumstances. They mount a detailed historical rebuttal of the Justice Department’s 1968 reading of the Insurrection Act’s interaction with election protection statutes, but acknowledge that constitutional theories like the “protective power” are murkier and remain untested in court. Either way, they warn that the legal ambiguity could enable a determined executive to deploy the military during elections before the courts could catch up.

Many commentators worry that a president might invoke the Insurrection Act to sidestep other statutory limits. That’s a reasonable concern. But the Insurrection Act is not the only mechanism for a president to deploy the military at or near a polling site in a manner at least arguably consistent with the law. And it may not be the most likely option for a president eager to use the troops. That’s particularly true because the interaction of these legal regimes—one permitting the president’s use of the military, the other restricting it—remains largely untested. A president with a propensity for pushing the limits of executive authority, including with respect to domestic deployments specifically, might be tempted to capitalize on the uncertainty by sending in the troops first and letting legal challenges follow.

Congress Should Do Something: The Case for (Fixing) the Great American AI Act

Charlie Bullock argues that the Great American AI Act, a bipartisan discussion draft crafted by Reps. Obernolte and Trahan, offers the best federal frontier AI safety framework proposed to date—but its limited whistleblower protections and sweeping state law preemptions mean it does more harm than good in its current form.

Perhaps the most important institutional advantage that GAAIA leverages is the capacity of federal agencies such as the Department of Commerce to carry out sophisticated rulemaking, a capacity built over decades of administering complex regulatory programs that no state agency can realistically match. GAAIA grants CAISI and the Department of Commerce broad authority to issue regulations fleshing out the auditing and transparency regimes outlined in Sections 111-112. Rulemaking! That word may not sound like the most exciting thing you’ve heard this week, but take my word for it: This is the good stuff.

The Unbearable Lightness of a Todd Blanche Confirmation Fight

Eric Columbus unpacks why Todd Blanche is likely to continue serving as acting attorney general despite not being confirmed by Congress. Columbus analyzes how the Attorney General Succession Act’s provision for the deputy attorney general to “exercise all the duties of that office” in case of a vacancy has been interpreted by courts to override the Federal Vacancies Reform Act and the Constitution’s Appointments Clause, allowing Blanche to serve indefinitely in an acting capacity.

Those who closely follow the frequent vacancy drama of the Trump years may wonder how this squares with the Federal Vacancies Reform Act of 1998 (FVRA). That act imposes time limits on how long actings can serve. In some cases, however, the FVRA takes a back seat. This is such a case.

But how long can Blanche serve under § 508? The text of the act—unlike the text of the FVRA—sets forth no limit. Can Trump really use it as an end run around the confirmation process? Unlike in many Trump-era legal puzzles, we have a recent, non-Trump precedent for this.

The Kill Switch and the Long Arm

Pablo Chavez considers concerns raised by European allies about reliance on U.S. technologies, specifically that the U.S. could cut off access or collect European users’ data. Washington, Chavez argues, should take action to protect European users that private companies or Europe cannot by implementing statutory constraints on the U.S. government’s own ability to interfere with U.S. companies’ services abroad.

Over time, the terms—if they function properly—should move into law. The power to cut off an ally or take its data sits mainly in permanent laws like the CLOUD Act, and the sanctions and export control statutes that only Congress can repeal. A promise to use that power sparingly isn’t equally permanent if it exists only in a program or an executive order, since any administration can erase it overnight while keeping the underlying power intact. That’s an asymmetry—permanent power, temporary restraint—and a key reason the restraint has to be set in a form as hard to undo as the power it checks.

Faithful Execution and the Removal Power

Nick Bednar and Todd Phillips highlight how the sweeping removal powers affirmed for the president in the Supreme Court’s decision in Slaughter may be in tension with the Take Care Clause of the Constitution. They note that some removals could leave multimember commissions below the threshold their statutes require for actions, such as with the recent firings at the Election Assistance Commission. In such cases, Bednar and Phillips argue that the commissions would be unable to execute their statutory obligations, thereby making the removals unconstitutional.

We propose a simple, objective test: Does the removal leave the commission below the threshold its statute sets for transacting business? If the answer is ”yes,” the removal is unconstitutional until it can be done without inhibiting the government’s functioning. This test would prevent multimember commissions, such as the Merit Systems Protection Board, from losing their quorums due to presidential removals alone. The president is free to nominate a replacement commissioner and remove the incumbent as soon as the Senate confirms the replacement. Allowing an alternative result violates the president’s obligations under the Take Care Clause.

From How Africa Works to How It Will Transform

Jane Munga reviews Joe Studwell’s book “How Africa Works: Success and Failure on the World’s Last Developmental Frontier.” Munga appreciates the nuance and insight Studwell brought to explaining Africa’s relative poverty through the continent’s demographic history and “low budget” colonialism, moving away from traditional conflict narratives. However, Munga finds that the book stumbles in its account of potential future economic transformation: digital technology appears throughout the book but is never treated as a central theme.

This is history that most African classrooms do not teach—an indictment of African education in itself, and one that leaves an African reader like me wondering whether our classrooms are doing enough to equip a new generation of problem solvers for Africa’s particular problems. The book’s achievement is to thread Africa’s history into one coherent account of why a resource-rich continent failed to emerge as an economic and political force when others did, and why the usual culprits explain so little—a diagnosis that now needs to travel both into classrooms and into policy offices. Studwell offers insight into questions analysts and governments still grapple with: why literacy remains low in most African countries (the burden, among others, of learning in languages not spoken at home); why short-termism in national economic assets persists; and what the lived legacies of colonialism, from patrimonialism to the resource curse, still are.

Supreme Court Undermines Section 702

In the latest edition of the Seriously Risky Business cybersecurity newsletter, Tom Uren discusses the implications of the Supreme Court’s decision in Slaughter for the Foreign Intelligence Surveillance Act Section 702, the increasing offensive cyber capabilities of Canada’s signals intelligence agency, and more.

When it comes to intelligence oversight, Trump dismissed three Democratic members of the Privacy and Civil Liberties Board (PCLOB) in January 2025, leaving it with just a single Republican member and consequently without a quorum. The PCLOB oversees the Data Protection Review Court and is one of the *ahem* “independent bodies” that the European Commission cited in its decision to approve the DPF.

Couple the Trump administration’s actions with the Supreme Court’s recent decision and we can see there might be reasons for concern. Schrems will also argue that this is a constitutional clash: EU treaty law demands independent supervisory authorities, but the U.S. Constitution now prohibits them.

Podcasts

Lawfare Daily: The Trials of the Trump Administration, July 2: I sit down with Eric Columbus, Kate Klonick, Molly Roberts, and Roger Parloff to discuss the Supreme Court’s rulings in the birthright citizenship case and Slaughter, indictments over purported vandalism at the Reflecting Pool, former CIA Director John Brennan’s civil suit against the Department of Justice, and geofencing warrants.

Scaling Laws, Founders & Founders: Kal Clark of Zauron Labs: Kal Clark joins Kevin Frazier to explain how AI is already reducing diagnostic errors in radiology. They talk about how Clark’s company’s AI system functions as a second-look safety layer, the scale of diagnostic error in modern healthcare, the institutional barriers that have prevented systematic second reviews, and privacy concerns around patient data.

Lawfare Daily: What’s Happening at ODNI?: Natalie Orpett talks with Feinberg and Curlee about ODNI, which was created to oversee the intelligence community but remains mysterious to the general public. They discussed what ODNI does, why it exists at all, and how recent developments are undermining its mission.

Scaling Laws: All Things Data Centers with Andy Masley: Andy Masley joins Kevin Frazier and Alan Rozenshtein to unpack the increasingly contentious issues related to data centers. They talk about Andy’s efforts to challenge what he sees as misleading claims about data center water use, land use, electricity rates, and local community impacts.

Lawfare Daily: The Military, Elections, and the Law: I talk with Orpett, Voss, and Roberts about the limits the Constitution and statutes put on the use of military in U.S. elections—as well as the arguments an eager executive might make to skirt those restrictions.

Lawfare Daily: Nuclear Weapons in the Age of AI, with Joshua Keating: Scott R. Anderson sits down with reporter Joshua Keating to discuss his new series on how artificial intelligence (AI) is impacting the use and development of nuclear weapons.

Rational Security: The “Scoot Over” Edition: Anderson sits down with me, Anna Bower, and Michael Feinberg to talk through the week’s big national security news stories, including last week’s Supreme Court decisions on executive power, firings at the Office of the Director of National Intelligence, and the supposed reconciliation between President Trump and his former lawyer, Michael Cohen.

Lawfare Daily: Prophecy, Prediction, and Power with Carissa Véliz: Tyler McBrien sits down with Carissa Véliz to discuss her new book, “Prophecy: Prediction, Power, and the Fight for the Future, from Ancient Oracles to AI.” They consider the history of prediction, why a healthy democracy—and a life well lived—requires uncertainty, and Véliz’s belief that “artificial intelligence is the new Oracle of Delphi and tech executives the new prophets.”

Scaling Laws, Founders & Founders: Adi Tantravahi of Cofactor: Adi Tantravahi, founder of Cofactor, joins Kevin Frazier to talk about how artificial intelligence (AI) can help cut through the administrative red tape that underlies and, in many ways, undermines the modern healthcare system.

Videos

I sit down with Anna Bower, Molly Roberts, Eric Columbus, and Roger Parloff to discuss the Justice Department settling a second suit with Michael Flynn, developments in the E. Jean Carroll litigation, the D.C. Circuit denying a stay pending appeal of the order to take Trump’s name off the Kennedy Center, and more.

Announcements

Lawfare is hiring a new Associate Editor, two new Data Journalism Fellows, and a Fall 2026 intern. Learn how to apply here, here, and here.

---

Today’s #BeastOfTheDay is the rabbit, seen here deciding not to attend a baseball game after all:

Video Source

In honor of today’s Beast, don’t be afraid to leave if you don’t like the vibe.