U.S. AI Safety Institute director leaves role

 The top official at the US AI Safety Institute is stepping down, raising new uncertainty about the future of a key government group focused on artificial intelligence under the Trump administration.

Elizabeth Kelly, the director of the safety institute and a face of US AI policy on the world stage, is set to leave her position by the end of the week, according to people familiar with the matter, who spoke on condition of anonymity to discuss private information.

Kelly and the safety institute did not immediately respond to a request for comment.

Housed under the Commerce Department, the US AI Safety Institute works with academics and developers to identify and mitigate risks from cutting-edge AI systems. It was created after former President Joe Biden signed an executive order in 2023 calling for leading AI companies to share safety test results and other critical information with the federal government.

The Biden administration tapped Kelly, a former White House aide, to lead the institute nearly a year ago. Her departure comes as world leaders and tech executives head to Paris for an AI summit to discuss the future of the technology. Vice President JD Vance will attend the event, according to a person familiar with the matter. Kelly and the safety institute helped host a similar gathering in San Francisco in November, shortly after Trump won the election.

Since taking office, Trump has moved swiftly to put his own stamp on AI policy, rescinding Biden’s order and calling for a new approach to boost US dominance in artificial intelligence. Trump also touted a massive new venture from OpenAI, SoftBank Group Corp. and Oracle Corp. that promises to invest $100 billion in infrastructure to support AI development.

But many in Washington, including Republicans, have commended the institute’s efforts to work with the tech industry to create voluntary standards on safe and responsible use for AI. Lawmakers on Capitol Hill also sought to advance legislation in the last Congress that would have formally authorized the institute. It employs more than two dozen staffers across offices in Washington and San Francisco.

“Elizabeth was an extraordinary leader for the Safety Institute at a fraught moment when success was far from preordained,” said Ben Buchanan, a former White House AI advisor under the Biden administration. “She built a startup that got things done within the bureaucracy, and recruited a world-class team to execute voluntary national security-focused AI testing and research.”

https://www.yahoo.com/news/head-us-ai-safety-institute-223226953.html