Anthropic Briefs House Homeland Security on AI in Closed-Door Session

Anthropic co-founder Jack Clark met behind closed doors Wednesday with members of the House Homeland Security Committee for a bipartisan briefing focused on national-security issues around artificial intelligence, Axios reports. According to multiple congressional sources, the discussion centered on technical topics such as model distillation—shrinking powerful AI systems into smaller versions—and U.S. export controls, while only briefly touching on Anthropic’s ongoing lawsuit challenging the Pentagon’s designation of the firm as a supply chain risk. Committee spokesperson Anna Holland said this session was part of a broader series of private roundtables with industry aimed at how the Department of Homeland Security evaluates, acquires and integrates emerging technologies like AI to strengthen critical infrastructure and cybersecurity. The meeting followed a December public hearing on AI and cybersecurity where companies including Anthropic and Google sent lower-level executives instead of CEOs, with leadership-level discussions shifted into these mostly private formats. The "friendly" tone described by participants highlights how much of Congress’ early AI security work—especially around export controls and government procurement—is occurring out of public view, raising transparency questions even as lawmakers try to get ahead of the technology’s risks.