U.S. AI safety chief resigns as policy scrutiny intensifies

Chris Fall resigned as director of the Commerce Department’s Center for AI Standards and Innovation after just three months in the job, leaving one of Washington’s few federal AI oversight hubs without a permanent leader as policymakers press for tougher scrutiny of advanced models. The department said a new director will be announced in the coming weeks, and Arvind Raman, who oversees the institute’s office at Commerce, will serve in the interim.

The vacancy matters because CAISI is not a symbolic post. National Institute of Standards and Technology materials say the center is the primary point of contact within the U.S. government for industry on AI testing and collaborative research, which puts it at the center of model evaluations, technical standards and discussions over how to measure frontier risks. The institute’s active docket includes the AI Agent Standards Initiative, announced on Feb. 17, 2026, and a Jan. 8 request for information on security considerations for AI agents.

CAISI also signed agreements on May 5, 2026, with Google DeepMind, Microsoft and xAI for frontier AI national security testing, a sign that the institute is trying to keep pace with rapidly advancing systems while building working relationships with the biggest developers. A leadership change at this point could slow coordination on those efforts just as companies are pushing more capable models into health care, finance, education, government services and national security uses.

Fall’s departure follows another turnover at the same institution. The U.S. AI Safety Institute’s director left the role on Feb. 6, 2025, and the institute was renamed CAISI in June 2025 as the Trump administration reoriented it toward standards and innovation under Commerce Secretary Howard Lutnick. That history has sharpened questions about continuity inside an agency that is meant to translate safety rhetoric into technical practice.

The broader policy backdrop leaves CAISI with outsized importance. The United States still lacks a single comprehensive federal AI law, so Commerce, NIST and sector regulators are filling the gap with guidance, requests for information and voluntary testing frameworks. With a permanent successor not yet named, industry groups, standards bodies and consumer advocates will be watching whether CAISI’s testing and standards agenda keeps moving or loses momentum during the transition.