Chris Fall, the head of the Center for AI Standards and Innovation (CAISI), has stepped down, as confirmed by the agency to several news organizations.

He took office only three months prior after the former appointee, Collin Burns, departed in under a week, The Washington Post reported then. Burns was said to have been “pushed out” of his role in April due to his former association with Anthropic amid tensions between the Trump administration and the company, sources informed the Post.

No explanation was provided for Fall’s exit. Before heading CAISI, Fall was the director of the Department of Energy’s Office of Science during Trump’s first term and had served as the acting director of the DOE’s Advanced Research Projects Agency-Energy. His previous experience includes work in the DOE’s Office of Naval Research (ONR).

Before Burns and Fall, the organization was overseen by venture capitalist David Sacks, who held the role of White House AI and crypto czar at that time. Sacks resigned in March.

CAISI, which functions under the National Institute of Standards and Technology, serves as the main body for establishing technical standards and assessment methods for AI models, as well as evaluating cybersecurity threats. However, it was not the agency involved in the latest model-risk controversy.

This controversy arose in June when the U.S. Commerce Department invoked a little-known export control directive that effectively compelled Anthropic to withdraw its Mythos and Fable models from the market. The restriction was lifted by the month’s end, with Secretary of Commerce Howard Lutnick expressing satisfaction with Anthropic’s safety strategies.

Earlier this month, the White House additionally endorsed an executive order for a new AI safety supervision initiative dubbed “Gold Eagle,” which establishes a clearinghouse for coordinating cybersecurity vulnerabilities. Numerous federal bodies were included in the program, such as the Commerce Department and the Department of Homeland Security. However, as CNBC highlighted, CAISI was not among the federal entities listed.

Meanwhile, after the ban on Anthropic’s models was lifted, Google DeepMind CEO Demis Hassabis began advocating for the establishment of an independent, industry-led standards organization to oversee frontier AI, similar to FINRA — essentially the same objective that CAISI was designed to address.

Fall’s departure also comes following this weekend’s concerns regarding the Chinese AI lab Moonshot’s latest version of its open model Kimi, which competed effectively against leading frontier models. The administration was contemplating measures to potentially restrict Chinese open models, Axios reported. This ignited immediate discussion and backlash over the weekend, including from Sacks, who contended that regulations should not be used as a means of protectionism for U.S. proprietary AI laboratories.

While CAISI has published a few reports assessing the capabilities of Chinese open-weight models like Z.ai’s GLM-5.2 and DeepSeek V4 Pro, it has been reticent regarding its testing procedures. (Open weight indicates that these models can be downloaded and operated locally, but their training code and datasets remain unavailable). Since July 9, TechCrunch has made several inquiries to both the DoC and NIST concerning the operations of its LLM evaluations and has not received any feedback.