Chris Fall has resigned as director of the U.S. Center for AI Standards and Innovation, or CAISI, after just three months in the post. CAISI confirmed the departure to multiple media outlets on Monday. The Commerce Department did not explain why he left and said only that Fall’s appointment had always been temporary.
Even by recent CAISI standards, three months counts as a relatively long stay. Collin Burns, who took over the unit in April, was removed in less than a week. The article says that decision was tied to his previous employment at Anthropic, at a time when the Trump administration was clashing with the company.
Three leaders in a little over six months
Before Burns, the first person overseeing CAISI was venture capitalist David Sacks, who held the title of White House AI and crypto czar. Sacks stepped down in March, and that title has not been filled since. Over a little more than six months, the U.S. government has now cycled through three people in a key AI oversight role. National Institute of Standards and Technology Director Arvind Raman is currently serving in an acting capacity.
CAISI sits under the National Institute of Standards and Technology, or NIST. It was previously the U.S. AI Safety Institute during the Biden era, then reorganized into the federal government’s main body for setting technical standards for AI models, running capability tests, and identifying cybersecurity risks. As described in the source, frontier AI labs would in theory need to clear this body’s safety review before releasing advanced models.
Before taking the CAISI job, Fall served as head of the Energy Department’s Office of Science during Trump’s first term. He also led the department’s ARPA-E on an acting basis and earlier worked at the Office of Naval Research. The Commerce Department said a formal successor would be announced in the coming weeks.
Major AI oversight moves have happened without CAISI
The report argues that CAISI has been largely absent from the most closely watched U.S. AI regulatory actions of the past six months.
In June, the Commerce Department invoked an obscure export-control provision and at one point forced Anthropic to pull two models, Mythos and Fable. The move came from Commerce Secretary Howard Lutnick, not CAISI. By the end of the month, Lutnick was satisfied with Anthropic’s safety plan and the restriction was lifted. CAISI was not visibly involved in that episode either.
Earlier this month, the White House signed an executive order creating a cybersecurity vulnerability coordination mechanism called Gold Eagle. Responsibility was assigned to the Commerce Department and the Department of Homeland Security. CAISI was again left out.
Industry calls are shifting outside government
After the Anthropic restrictions were reversed, Google DeepMind CEO Demis Hassabis called for a frontier AI standards body modeled on FINRA, the self-regulatory organization in the securities industry. He said it should be independent of government and led by industry participants. According to the source, that overlaps with the mission CAISI was originally supposed to carry out.
Debate over Chinese open-source models adds pressure
Just before Fall resigned, Moonshot released a new version of its open-source Kimi model. The article says its benchmark performance caught up with mainstream frontier systems, prompting fresh debate inside the U.S. government over whether Chinese open-source models should be banned from use.
Sacks pushed back in that debate, arguing that regulation should not be used as a trade tool to protect private U.S. model developers.
CAISI has had little to show in that fight. The agency has published several evaluation reports on Chinese open-weight models, including Z.ai’s GLM-5.2 and DeepSeek V4 Pro. But the source says it has never publicly explained how those tests were run or how the standards themselves were set.