Congress Punts on AI Safety Regulation as Labs Beg for Someone to Hit the Brakes

House Speaker Mike Johnson signaled Sunday that Congress won't take the lead on regulating AI safety, even as pressure from inside the industry itself reached an unusual crescendo. Johnson, appearing on both CNN and NBC's Meet the Press, acknowledged he'd consider recalling House members from recess and scheduling a vote — but only if lawmakers already had a concrete solution ready to pass, a threshold that doesn't currently exist.

Instead of legislation, Johnson floated a softer approach: organizing a summit between President Trump, members of Congress, and the chief executives of the country's leading AI companies. The proposal reflects a familiar Washington instinct to convene rather than legislate, particularly on a technically complex issue where political consensus is still forming.

The backdrop makes the inaction notable. This week, the four largest AI labs — a group that includes the companies spending billions racing each other to the frontier — endorsed slowing their own models' development, a remarkable case of industry self-restraint that typically only emerges when companies sense regulation is coming regardless. The moment was crystallized by a 3,800-word essay published Saturday by Anthropic CEO Dario Amodei, in which he argued for deliberately decelerating AI development and layering in more rigorous safety measures. Several other AI leaders publicly aligned with Amodei's position.

On the Democratic side, House Minority Leader Hakeem Jeffries told ABC's This Week that Democrats were planning to huddle Tuesday morning to discuss AI — suggesting the minority party may move faster than Republican leadership in at least articulating a legislative position. Whether that translates to any actual floor action depends heavily on Johnson, who controls the House calendar.