House Lawmakers Form 26-Member AI Safety Caucus
The bipartisan group plans to propose safety legislation and convene briefings after high-profile security incidents involving AI systems.

A bipartisan group of 26 House lawmakers has formed a caucus to develop targeted artificial intelligence safety policy and help Congress respond to high-profile security incidents involving AI systems.
Reps. George Whitesides, D-Calif., Nathanial Moran, R-Texas, and Don Beyer, D-Va., are co-chairing the AI Safety and Innovation Caucus. Seven of its 26 members are Republicans.
The caucus plans to propose AI safety legislation and bring developers, independent experts and government officials together to brief lawmakers after prominent security incidents. It also aims to connect members with outside experts and give lawmakers outside relevant committees tools to participate in policy discussions.
Whitesides said the group wants to help refine responsible legislation for consideration on the House floor. Moran said the caucus would seek consensus and targeted policies that protect people while allowing innovation.
The lawmakers cited a July hack of Hugging Face servers in which, they said, autonomous agents developed by OpenAI broke out of a sandbox environment. The lawmakers also cited an Anthropic incident in which the company said its agents escaped testing environments and accessed third-party systems.
The caucus says it intends to pursue safeguards while preserving innovation and U.S. leadership in the AI industry.