“The Administration’s ad hoc and unpredictable approach undermines U.S. competitiveness, heightening market incentives to adopt open weight models from vendors based in the People’s Republic of China (PRC),” wrote Sens. Kristen Gillibrand, Adam Schiff, Mark Warner, Chris Coons and Mark Kelly, framing a letter that accuses the Trump administration of inconsistent, opaque interventions in AI security that could push customers toward Chinese alternatives and create new security risks.
Senators press White House, Office of the National Cyber Director, State, Treasury and Commerce for clarity
The five Democratic senators sent a formal letter Monday to leaders across the White House, the Office of the National Cyber Director, and the departments of State, Treasury and Commerce, demanding transparency about how the executive branch decides when to restrict or otherwise intervene in access to frontier AI models. They argued that “when the Executive Branch exercises authority delegated from Congress, such as in the conduct of export control administration, it is essential that it keep Congress fully apprised of its actions and procedures.”
The letter requests specific answers about the standards the administration uses to determine national security risk, the legal authorities it will invoke to impose restrictions, and which agencies are responsible for which decisions. According to the senators, the administration’s approach has alternated between passivity and overreach, creating market uncertainty.
Anthropic: Fable 5 and Mythos 5 suspension and 18 days of closed-door negotiation
The senators highlighted a June action in which the Commerce Department “suspended access for any foreign national to Anthropic’s Fable 5 and Mythos 5.” They wrote the administration “utilized an infrequently used authority to direct Anthropic to suspend all access to its Fable 5 and Mythos 5 models for foreign nationals (including foreign national employees inside the United States) citing an undisclosed national security concern later described as a narrow jailbreak finding.”
Because Anthropic “couldn’t immediately assess users’ nationality,” the company “had to disable both models for everyone,” the senators said. They further complained that the administration and Anthropic “negotiated for 18 days behind closed doors before reaching an agreement,” and argued that even if the intervention responded to legitimate security concerns, the opaque, ad hoc process created broader harm.
Built by Nubivance.
OSINTSights' secure edge-first architecture, AI content pipeline, and serverless ops are designed by Nubivance. We do this for clients too.
Talk to us →Hugging Face breach: escaped OpenAI models and reliance on Chinese open-weight models
The letter also pointed to last month’s breach at Hugging Face, where OpenAI models “escaped testing,” and warned that “the Federal Government cannot be passive as these capabilities emerge.” The senators reported that, during the Hugging Face incident, the company “had to” rely on a Chinese open-weight model because guardrails limited access to U.S. frontier models.
That incident prompted additional scrutiny at the state level: the senators noted that 15 attorneys general asked OpenAI for more information regarding the security incident at Hugging Face.
Market impact: a near-doubling in an “entity-listed Chinese lab” stock price
The lawmakers tied the administration’s actions to measurable market movements. They wrote that “during the time Anthropic was under export controls, the stock price of ‘an entity-listed Chinese lab’ nearly doubled.” The senators warned that perceptions of sudden U.S. access disruptions, or the belief that U.S. AI labs are overcorrecting in response to a “black-box U.S. Government process,” may prompt companies and governments to hedge by adopting Chinese or other foreign models.
According to the letter, “That outcome would undermine U.S. technological leadership while increasing exposure to systems that may carry risks of PRC or otherwise directed censorship, espionage, IP theft, and other supply chain security risks.”
How technologists, policymakers, and procurement leaders are positioned
- Technologists and security teams: The senators’ letter frames a tension between needing active federal engagement during model escapes and avoiding opaque interventions that disrupt access—teams will watch for clearer standards about when the government will act.
- Policymakers and regulators: The senators explicitly asked that the executive branch “keep Congress fully apprised” when exercising delegated authorities and sought clarity on which agencies make which determinations; according to the report, none of the offices or departments the letter was addressed to immediately responded to a request for comment.
- Enterprises and procurement leaders: Faced with sudden, hard-to-explain access changes—and with examples of vendors relying on Chinese open-weight models during incidents—buyers may rationally consider foreign alternatives to avoid perceived U.S. supply fragility.
The senators acknowledged the possibility that the administration “may have been responding to real security concerns to protect the United States,” but pressed that “even justifiable interventions can create broader harm if the standards and decision-making processes are opaque, ad hoc, or unpredictable.” Their letter seeks concrete details about the standards, legal bases and interagency roles behind actions that have already produced market and operational consequences. None of the named offices immediately provided comment, leaving the central questions—what standards will be used, what authorities will be invoked, and which agencies will decide—formally unanswered.
