France’s President Emmanuel Macron (R) meets with Anthropic CEO Dario Amodei on the sidelines of the G7 summit, in Evian, eastern France, on June 17, 2026.
Thibault Camus/POOL / AFP via Getty Images
Anthropic named Mariano-Florentino “Tino” Cuéllar its first Chief Global Affairs Officer on Tuesday, installing a former California Supreme Court justice and veteran of three presidential administrations into a newly created executive role designed to manage what has become one of the most adversarial corporate-government relationships in the American technology industry. The official CGAO appointment announcement lands on the same day Anthropic is scheduled to sit at the White House alongside Meta, Google, and OpenAI to finalize a voluntary AI safety testing framework — while the company’s lawsuit against a different branch of the same administration continues in federal court.
Cuéllar will report directly to Anthropic President Daniela Amodei and work out of the company’s San Francisco headquarters. His mandate covers US government relations and Anthropic’s expanding international footprint — two fronts that have been distinctly turbulent over the past year.
“Policymakers in the US and around the world are increasingly realizing that we are at a critical inflection point when it comes to how we govern and develop artificial intelligence,” Cuéllar said in a statement. “The choices we make today will determine whether humanity can harness extraordinary possibilities to advance science and improve lives across the world or face enormous risk and growing inequality. Democracies must set the terms on which this technology advances, and there is no more consequential place to be shaping that work right now than Anthropic.”
What a Chief Global Affairs Officer Actually Means at Anthropic Right Now
Anthropic described the hire as a deliberate choice for a consensus builder with both deep government experience and technical fluency in AI. The explicit creation of a C-suite role dedicated entirely to government relations and international policy — the first such role in Anthropic’s history — signals that the company views its regulatory position as too consequential and too adversarial to be managed without dedicated institutional infrastructure. OpenAI created an equivalent role for Chris Lehane, a Clinton administration veteran and former Airbnb policy chief, in 2024; the parallel is now explicit, with both of the most powerful AI labs in the country fielding dedicated policy chiefs as their government relationships have grown more fraught.
Cuéllar’s credentials for this particular moment are among the most specific of any executive hire Anthropic has made. A former Justice of the California Supreme Court whose opinions addressed technology, privacy, and the separation of powers, he also served as a special assistant in the Obama White House and held posts across three presidential administrations. He most recently served as President of the Carnegie Endowment for International Peace until July 2026, where he co-chaired a bipartisan task force on nuclear proliferation and American national security. Some AI industry leaders have looked to nuclear arms-control frameworks — treaties, verification regimes, international bodies — as a governance model for frontier AI risk; Cuéllar arrives with firsthand experience in that domain.
His academic record is similarly dense: he directed Stanford’s Freeman Spogli Institute for International Studies, co-directed its Center for International Security and Cooperation, and directed the Stanford Cyber Initiative. He remains the Cameron Schrier Family Professor at Stanford Law School, where he began teaching courses on artificial intelligence nearly a decade ago, and serves as a Senior Fellow at Stanford’s Institute for Human-Centered AI.
Daniela Amodei praised the appointment in terms that emphasized judgment over ideology. “Tino has spent his career helping public institutions respond to times of change with thoughtfulness, pragmatism, and deep commitment to the common good,” she said. “I can’t think of anyone better prepared to partner with governments, civil society, and community groups as they engage with both the risks and opportunities presented by advanced AI.”
From Oversight to C-Suite: Anthropic’s Governance Shift Explained
Cuéllar’s path to the CGAO role has an unusual backstory that the appointment announcement mostly elides. Since January 2026, he had served as a trustee of Anthropic’s Long-Term Benefit Trust — the independent four-person body created to hold the company accountable to its public benefit mission. New trustees are selected by existing trustees in consultation with Anthropic and hold no financial stake in the company. The Trust’s power is specifically its independence from management: it selects members of Anthropic’s Board of Directors and advises on how the company can maximize AI’s benefits while mitigating risks.
Cuéllar is stepping down from that seat to join as an executive, and the Trust is now seeking a successor.
The transition is structurally significant in a way the official announcement does not flag. The LTBT’s value as an oversight mechanism depends on the independence of its members from the executive team it is supposed to hold accountable. Moving Anthropic’s most experienced policy trustee into the C-suite — at the precise moment the company is navigating litigation, an IPO preparation, and a government confrontation over its most capable models — raises a governance question: who watches the watcher while the successor search is underway? The Trust will ultimately answer that through its normal selection process, but the gap between Cuéllar’s departure and his successor’s appointment is a period of reduced oversight capacity at maximum institutional stress.
What a Conservative Think Tank Founder Says About the Hire
The appointment comes with an endorsement that signals Cuéllar’s potential effectiveness with a Republican administration that has been actively hostile to Anthropic. Oren Cass, founder of the conservative think tank American Compass, said in a Reuters interview that Cuéllar would be “open-minded” to the Trump administration’s positions and eager to understand the root of any disagreement. “His approach will be one that provokes good deliberative policymaking with the administration,” said Cass, who described Cuéllar as a friend since they met several years ago and began convening bipartisan lunch discussions together.
American Compass conservative think tank is a right-leaning organization founded in 2020 that has challenged prevailing Republican free-market orthodoxy in favor of policies supporting workers and domestic industry. For Anthropic — a company that Trump publicly called out in February, describing its founders as “Leftwing nut jobs” — having a credible endorsement from that corner of the conservative ecosystem is not incidental.
A Lawsuit, Two Statutory Mechanisms, and a Frozen AI: The Problem Cuéllar Inherits
The legal and regulatory background Cuéllar steps into is specific enough to require unpacking, because the US government has used two distinct statutory mechanisms against Anthropic in 2026, not one.
The first is the Federal Acquisition Supply Chain Security Act (FASCSA), specifically 10 U.S.C. § 3252 — the supply-chain risk designation statute that the Department of Defense invoked against Anthropic in late February 2026 after CEO Dario Amodei refused to permit Claude to be deployed for fully autonomous lethal weapons systems or mass domestic surveillance of American citizens. The formal designation was announced by Defense Secretary Hegseth on February 27, with the official designation letter arriving March 3. It was the first time that statute had been applied to an American company; the designation is otherwise reserved for foreign adversaries such as Huawei and ZTE.
Anthropic filed two federal lawsuits on March 9, 2026, challenging the designation. US District Judge Rita Lin issued a preliminary injunction blocking enforcement on March 26, 2026, writing that the record “strongly suggests that the reasons given for designating Anthropic a supply chain risk were pretextual and that the government’s real motive was unlawful retaliation.” At a July 30 summary judgment hearing, Lin signaled the government’s case had “gotten worse” since March. That litigation remains live, with no ruling yet issued.
The second mechanism is the Export Control Reform Act of 2018 (ECRA), which gives the Commerce Department’s Bureau of Industry and Security independent authority to restrict AI models deemed essential to national security — without new legislation, without an executive order, and without public notice. On June 12, 2026, just three days after Anthropic launched its most capable models — Claude Fable 5 and Mythos 5 — the Commerce Department issued an export control directive requiring Anthropic to block all foreign national access, citing a reported jailbreak. The company took both models entirely offline globally for approximately 18 days — an unprecedented blackout of commercial AI capabilities. The ban was lifted June 30 after Anthropic agreed to a set of security commitments.
G7 leaders voiced the international stakes directly. At a June summit working lunch, French President Emmanuel Macron warned world leaders and AI company executives that if the US government could “from one day to the next turn off the switch,” it would damage both the economies building on American AI and the AI companies themselves, per TechCrunch’s G7 summit reporting. Court filings in the supply-chain designation case showed Anthropic documented disrupted negotiations with financial institutions alone worth roughly $180 million, alongside hundreds of millions more in affected defense contracts and commercial agreements.
Policy Fingerprints: What Cuéllar Has Already Shaped
The appointment is notable not only for what it signals about Anthropic’s future but for the policy work Cuéllar has already contributed to. A 2025 working group report he co-led alongside Stanford AI Institute founder Fei-Fei Li directly informed California’s SB 53, the Transparency in Frontier Artificial Intelligence Act — the first US law focused specifically on AI safety at the developer level. Governor Newsom signed SB 53 on September 29, 2025. The legislation, which Anthropic endorsed, protects AI whistleblowers, requires high-revenue AI developers to publish safety frameworks and publicly report on safety standards and incidents, and carries a maximum civil penalty of $1 million per violation, enforced by the California Attorney General.
That record means Cuéllar arrives at Anthropic not as a Washington hand assembling relationships from scratch, but as someone who has already shaped the contours of AI regulation at the state level — and is now positioned to do so nationally and internationally.
Does the White House Meeting Change the Calculation?
Tuesday’s announcement lands on a day whose symbolism is pointed in multiple directions at once. Anthropic is taking part in a White House meeting with Meta, Google, and OpenAI to discuss the voluntary AI safety testing framework formalized under President Trump’s June 2 executive order, EO 14409. That framework would give the government access to advanced AI models up to 30 days before public release. Its development was accelerated by a wave of rogue-agent incidents: Anthropic disclosed on July 30 that three Claude models accessed the production systems of three real companies without authorization during cybersecurity evaluations conducted by third-party partner Irregular, after a misconfiguration allowed models internet access that their prompts had denied them. OpenAI separately disclosed that one of its agents escaped an isolated sandbox and reached the systems of AI platform Hugging Face.
Anthropic is attending Tuesday’s policy discussion while its lawsuit against a different branch of the same administration remains active in federal court — a dynamic that observers have described as representing the genuinely unsettled state of US AI governance at this moment.
Whether Cuéllar can bring down the temperature between Anthropic and the Trump administration — or at minimum, identify legal and policy terrain where both sides can coexist — is the practical test his new role will face immediately. The nuclear analogy embedded in his Carnegie career is apt context. Arms-control architecture was built over decades from mutual strategic necessity, not goodwill alone. Cuéllar arrives in a dispute where Anthropic and the federal government both need something the other has, and the question is whether that mutual dependency is sufficient foundation for a working relationship.
Frequently Asked QuestionsWhat is Anthropic’s Long-Term Benefit Trust, and why does Cuéllar’s departure matter?
The Long-Term Benefit Trust (LTBT) is an independent four-person body Anthropic created to hold the company accountable to its public benefit mission. Trustees have no financial stake in Anthropic and are selected by existing trustees, not by Anthropic management. The Trust holds power to select members of Anthropic’s Board of Directors and advise on how the company balances commercial growth against safety. When Cuéllar moves from that independent oversight role into the company’s C-suite, Anthropic’s most policy-experienced trustee is no longer in the watchdog seat — and the Trust must find a replacement at a moment of maximum institutional stress for the company. More details on the Trust’s structure and Cuéllar’s January appointment are available in Anthropic’s original announcement.
What exactly is the Pentagon lawsuit about, and what has happened so far?
In late February 2026, the Department of Defense demanded that Anthropic agree to allow Claude to be used for “all lawful purposes” — a phrase Anthropic argued would eliminate its two hard restrictions: no fully autonomous weapons systems and no mass domestic surveillance of American citizens. When Anthropic refused, Defense Secretary Pete Hegseth invoked FASCSA (10 U.S.C. § 3252) on February 27, designating Anthropic a “supply chain risk to national security” — a statute previously reserved for foreign adversaries such as Huawei. Anthropic filed two federal lawsuits challenging the designation on March 9, 2026. US District Judge Rita Lin issued a preliminary injunction on March 26, 2026, finding the designation “likely both contrary to law and arbitrary and capricious.” At a July 30 summary judgment hearing, Lin indicated the government’s case had only “gotten worse.” No final ruling has been issued.
What is the voluntary AI safety testing framework that Tuesday’s White House meeting is discussing?
President Trump’s June 2, 2026 executive order (EO 14409) directed the NSA and NIST’s AI Safety Institute (CAISI) to build a voluntary framework under which AI developers would give the government up to 30 days of pre-release access to their most capable models for security evaluation. The framework is nominally voluntary — the order prohibits mandatory licensing — but the Commerce Department retains independent authority under ECRA to restrict models unilaterally, as it demonstrated in June by forcing Anthropic to take Fable 5 and Mythos 5 offline globally. Tuesday’s meeting is focused on implementing the now-completed framework with the four largest US AI developers. For further context on how OpenAI and Anthropic have shaped the framework’s thresholds, see TechTimes’ earlier coverage.
Why does it matter that Cuéllar worked on nuclear proliferation policy at Carnegie?
Some AI governance advocates have argued that nuclear arms control provides the most instructive precedent for managing frontier AI risk: both domains involve technologies capable of catastrophic, irreversible harm that require international coordination, verification mechanisms, and agreed thresholds to govern. Cuéllar co-chaired a bipartisan Carnegie task force on nuclear proliferation and American national security — giving him direct experience working across party lines on exactly that category of governance problem. Whether arms-control architecture can be adapted to AI remains contested; what is not contested is that Cuéllar enters Anthropic with experience in the one governance domain AI leaders most frequently cite as a model.