Anthropic’s Jack Clark Meets with the U.S. House Committee on AI Model Distillation and Export Controls

According to 1M AI News, Jack Clark, co-founder and head of public interest at Anthropic, held a closed-door roundtable on March 19 with bipartisan members of the House Committee on Homeland Security. The discussions primarily focused on model distillation (compressing large AI systems into smaller versions) and export controls, with disputes over Pentagon contracts not being central to the meeting. Attendees described the atmosphere as "friendly."

This is the first of a series of closed-door roundtables between the committee and AI industry stakeholders, with additional sessions planned to cover more companies over the coming months. Committee spokesperson Anna Holland said the shift to a closed-door format was “to enable more substantive discussions,” focusing on critical infrastructure security, cybersecurity, and how the Department of Homeland Security evaluates and adopts new AI technologies. Clark recently assumed the new role of Head of Public Interest at Anthropic, as the company expands its government relations presence in Washington.

This meeting occurs amid ongoing controversy between Anthropic and the Pentagon. Earlier this month, Anthropic formally sued the Department of Defense to overturn the “supply chain risk” designation; this week, the Pentagon submitted documents to the court alleging security concerns due to its employment of a large number of Chinese nationals. Meanwhile, Anthropic continues to actively strengthen its relationships with congressional lawmakers.