Trump’s head of AI safety agency CAISI resigns after months on job

Chris Fall, head of the Center for Artificial Intelligence Standards and Innovation, resigned as director just three months after taking office. selected CNBC confirmed Monday that the Trump administration has accepted the job.

CAISI, part of the U.S. Department of Commerce, aims to help the government facilitate “testing and collaborative research” around commercial AI systems, according to its website.

Arvind Raman, director of the National Institute of Standards and Technology, will serve as CAISI’s acting director and continue to oversee the agency, Commerce Department spokeswoman Kristen Eichamer told CNBC. Eichamer did not explain why Fall, who was elected to lead the organization in April, resigned from his position.

axios He was the first to report Fall’s departure.

His exit adds to growing uncertainty about who serves as the Trump administration’s key figure in AI policy and research. Venture capitalist David Sacks previously served as artificial intelligence and cryptocurrency czar at the White House, but he left the post in March and has not yet been replaced.

This is a particularly sensitive moment for the White House as Chinese open-source models gain traction in the US against large proprietary models from OpenAI and Anthropic. Late last week, Chinese startup Moonshot AI unveiled its new model, the Kimi K3, which it says has closed the gap on leading US offerings and outperformed the most capable systems from OpenAI and Anthropic in some benchmarks.

The Trump administration has taken a more hands-on approach to AI regulations since the president signed an AI executive order in June. The order asks AI developers to voluntarily provide models to the government to evaluate their capabilities ahead of full release and gives federal agencies 60 days to develop an evaluation framework.

This process was dark for companies trying to launch powerful new models during this period.

OpenAI said in June that it had agreed to limit distribution of the GPT-5.6 model series to a group of “trusted partners” at the request of the US government. Two weeks ago, Anthropic was forced to disable access to Fable 5 and Mythos 5 models in order to comply with the export control directive issued by the Ministry of Commerce. Both companies were later able to release their models more widely.

The Trump administration has taken steps in recent weeks to begin enforcing the executive order; clearing house It is called “Golden Eagle”. The clearinghouse aims to find and fix cybersecurity vulnerabilities and, as CNBC previously reported, blames the White House for giving companies the green light to access the latest AI models.

The clearinghouse has already begun “receiving and prioritizing identified cyber vulnerabilities” and “coordinating scan verifications.” a version From the White House. CAISI was not named as an agency involved in its development.

WRISTWATCH: Fmr. Trump administration senior WH policy advisor on lifting restrictions on Anthropic’s AI models