US AI safety test chief quits after just 3 months

Chris Fall, head of the U.S. AI Safety Institute (AISIC), resigned just 3 months after taking the job. The center works with labs like Anthropic, Google DeepMind, OpenAI, and xAI to test AI models for safety flaws before release, aiming to prevent rivals from using AI for chemical or biological weapons.

A Commerce Department official said Arvind Raman, former dean of Purdue Engineering, will temporarily replace Fall. Fall's exit shows the U.S. government's AI policy keeps shifting. The Trump administration first promised less tech regulation, then moved to stricter AI oversight, with policy directions changing almost weekly.