The Trump administration’s effort to build stable guardrails for artificial intelligence is being tested by yet another abrupt leadership change at its top AI safety agency.

In April, former Department of Energy official Chris Fall was appointed to lead the Center for AI Standards and Innovation (CAISI), the federal hub for testing frontier AI models and developing technical standards.¹ The newly reorganized center, previously known as the U.S. AI Safety Institute, was tasked with evaluating advanced systems and supporting standards for safe deployment.²

Within weeks, Fall was drawn into tense policy battles. After the Commerce Department imposed export-style controls that forced Anthropic to pull its advanced Fable 5 and Mythos 5 models from the market in June, Fall “was closely involved in discussions with top Anthropic executives about how to restore access” to those systems.¹ The ban was lifted by the end of the month, when Commerce Secretary Howard Lutnick said he was satisfied with Anthropic’s safety plans.³

At the same time, the White House signed an executive order launching “Gold Eagle,” a new AI safety oversight program and cybersecurity vulnerability clearinghouse that named multiple federal agencies but pointedly did not include CAISI, raising questions about the center’s clout inside the administration.³

On July 20, Commerce officials confirmed that Fall had resigned just three months into the job, describing his exit as an “abrupt departure” even as they insisted his appointment was always meant to be temporary.² The move makes him the latest short‑tenured AI “czar” after Collin Burns, who left in under a week amid concerns over his prior work for Anthropic, and venture capitalist David Sacks, who stepped down in March.³

NIST Director Arvind Raman will serve as acting CAISI chief while leading the search for a permanent replacement, which Commerce says it hopes to announce in the coming weeks.¹ With the reasons for Fall’s departure unclear and AI policy priorities still shifting, the agency at the center of U.S. AI testing now faces renewed uncertainty about its role and direction.⁴