Chris Fall has resigned as director of the Center for AI Standards and Innovation (CAISI), the agency confirmed to multiple news outlets. His departure comes just three months after he was appointed, making him the third leader to leave the post in a short span.
Fall’s exit follows a turbulent period for the agency, which operates under the National Institute of Standards and Technology (NIST). No reason was given for his departure. Fall previously served as director of the Department of Energy Office of Science during the first Trump administration, was acting director of DOE ARPA-E, and worked at the DOE Office of Naval Research.
A revolving door at CAISI
Fall’s resignation is the latest in a series of leadership changes at CAISI. David Sacks, the former White House AI and crypto czar, stepped down in March. Collin Burns, the next appointee, left in less than a week. Burns was reportedly pushed out in April because he previously worked for Anthropic and the Trump administration had been battling with the company, sources told the Washington Post.
The agency, created to develop technical standards and testing methods for AI models and assess cybersecurity risks, has struggled to maintain stability. CAISI was not at the center of the most recent model-risk brouhaha, and it was not among the federal organizations mentioned in the Gold Eagle executive order signed earlier this month. Gold Eagle creates a clearinghouse for cybersecurity vulnerability coordination, with the Commerce Department and Department of Homeland Security named in the program.
Anthropic ban and aftermath
In June, the US Commerce Department invoked an obscure export control directive forcing Anthropic to pull its Mythos and Fable models. The ban was lifted by the end of June after Secretary of Commerce Howard Lutnick said he was satisfied with Anthropic’s safety plans. After the ban was lifted, Demis Hassabis, CEO of Google DeepMind, called for an independent, industry-run standards body for frontier AI modeled after FINRA.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Chinese open-weight models under scrutiny
Fall’s resignation follows handwringing over Moonshot, a Chinese AI lab that released a new version of its open model Kimi that performed competitively against flagship frontier models. The administration is weighing efforts to ban Chinese open models, Axios reported. Debate and outrage over the weekend included comments from Sacks, who argued that “regulations shouldn’t be used as protectionism for US proprietary AI labs.”
CAISI has released reports on the capabilities of Chinese open-weight models, including Z.ai’s GLM-5.2 and DeepSeek’s V4 Pro. Open-weight models can be publicly downloaded and run locally, but training code and datasets are not available. The agency has not talked much about its testing processes. Since July 9, TechCrunch has sent multiple inquiries to the Department of Commerce and NIST about LLM evaluations but has not received a response.
The road ahead for AI standards
The leadership turmoil at CAISI raises questions about the US government’s ability to set technical standards for AI. The agency was designed to be the primary organization for developing testing methods and assessing cybersecurity risks, but it has been overshadowed by other federal efforts. The Gold Eagle program, which does not mention CAISI, is the latest White House initiative for AI safety oversight.
With three directors in a matter of months, the agency faces an uncertain future. The administration’s focus on Chinese open models and the debate over protectionism versus regulation adds further complexity. Sacks’ weekend comments highlight the tension: “Regulations shouldn’t be used as protectionism for US proprietary AI labs.”
As the US weighs its next steps, the absence of a stable leadership at CAISI could slow progress on technical standards that many in the industry say are urgently needed.