Chris Fall resigns as head of US AI safety center after three months

Chris Fall has resigned as head of the U.S. Center for AI Standards and Innovation three months after being chosen for the role, CNBC reported Monday. The departure creates another leadership gap at a federal office charged with assessing frontier artificial intelligence systems for security risks before public release.

A Commerce Department spokesperson confirmed Fall’s resignation to CNBC but did not give a reason. CAISI sits within the Commerce Department and is linked to the National Institute of Standards and Technology, the federal standards agency known as NIST.

Arvind Raman, NIST’s director, will serve as acting director of CAISI while continuing to run NIST, CNBC reported, citing the Commerce Department spokesperson. That arrangement places the AI safety center under interim leadership at a time when the federal government is seeking earlier access to advanced AI models and more formal testing of their capabilities.

CAISI works with major AI developers on national security testing, according to Reuters. The agency’s responsibilities include testing unreleased models for vulnerabilities, assessing risks from advanced systems and reducing opportunities for U.S. adversaries to exploit AI, Reuters reported.

Those responsibilities have expanded in parallel with federal efforts to scrutinize frontier models before deployment. PYMNTS reported in May that Google DeepMind, Microsoft and xAI signed agreements with CAISI to share frontier AI models with the agency for national security testing before release. Anthropic and OpenAI had previously signed agreements with CAISI’s predecessor in August 2024 and later revised them to conform with CAISI’s directives, according to PYMNTS.

President Donald Trump signed an executive order in June asking companies to give the White House access to new AI models before release, participate in benchmarking to examine model capabilities and allow trusted government partners to use models to strengthen cybersecurity for critical infrastructure, PYMNTS reported.

Interim leadership at a central AI office

CAISI’s role gives it a technical function within Washington’s AI policy apparatus. Rather than writing broad rules alone, the center tests systems, coordinates with laboratories and helps assess whether advanced models could expose cyber, national security or misuse risks before they reach the market.

Fall’s exit follows earlier turnover at the same office. The Hill reported Monday that Collin Burns, Fall’s predecessor at CAISI, was removed by the White House four days after taking the job because of ties to Anthropic, a company that had clashed with the administration. The Hill also reported that Fall’s appointment was intended to be temporary.

CNBC reported that the White House did not name CAISI as a participant when it announced Gold Eagle, a clearinghouse focused on identifying and fixing cyber vulnerabilities. The omission drew attention because CAISI’s remit includes work at the intersection of AI systems, cybersecurity and national security testing.

Fall’s resignation also comes after a separate vacancy in the administration’s technology policy structure. Former White House AI and crypto czar David Sacks stepped down in March and has not been replaced, CNBC reported. The result, according to CNBC, is that the White House lacks a clear point person for AI policy and research.

For AI developers, CAISI remains the federal office named in agreements to review frontier systems before public release. For policymakers, Raman’s acting appointment keeps the center operating inside NIST while the administration decides how to fill the role on a longer-term basis.

This story draws on original reporting from PYMNTS.