Commerce Department hunts for new AI safety director as leadership turmoil continues

The US Department of Commerce is searching for a new director to lead its AI governance body, the latest chapter in a revolving door of leadership that has plagued the organization since its founding. The institute, originally called the US AI Safety Institute, has now been rebranded, restructured, and left without stable leadership.

A timeline of departures and disruptions

The US AI Safety Institute launched in 2024 under the Biden administration with a clear mandate: manage the risks of advanced AI systems. Elizabeth Kelly, its first director, held the role until early February 2025, when she stepped down.

Her departure kicked off a rapid series of personnel changes. A replacement was brought in with ties to Anthropic, the AI company behind Claude. He lasted four days before being shown the door over concerns about his prior work.

The instability didn’t stop at staffing. On July 31, 2025, Commerce Secretary Howard Lutnick announced the institute would be rebranded as the Center for AI Standards and Innovation, or CAISI. The word “safety” was deliberately dropped from the name, shifting the institute’s focus from regulatory oversight toward national security and global competitiveness.

The bigger picture: standards without a standard-bearer

The original AI Safety Institute was designed to be America’s answer to a global challenge. The EU has its AI Act. China has been rolling out its own AI regulations. The US was supposed to have a credible, technically sophisticated body that could shape international standards and negotiate with allies.

The rebranding to CAISI signals a shift in the current administration’s philosophy. By foregrounding “standards” and “innovation” over “safety,” the Commerce Department is indicating that the US plans to compete first and regulate second.

The Commerce Department hasn’t announced a timeline for filling the position.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.