A bipartisan group of 29 House lawmakers has launched the AI Safety and Innovation Caucus, an initiative created to strengthen federal oversight of advanced AI systems while supporting U.S. leadership in AI development.
The caucus will focus on developing AI safety legislation and coordinating congressional responses to major security incidents. The effort is a response to growing concern that autonomous AI agents could create major cybersecurity threats as their capabilities continue to grow.
The group is led by Representatives George Whitesides (D-Calif.), Don Beyer (D-Va.) and Nathaniel Moran (R-Texas). Its membership includes 21 Democrats and eight Republicans, a notable bipartisan coalition given the Trump administration’s opposition to extensive AI regulation.
A central goal is to move beyond congressional discussions about AI and develop legislation that can attract support from both parties. The caucus will develop proposals, evaluate existing bills and help lawmakers reach agreement on AI safeguards. It also plans to consult with tech developers and cybersecurity specialists to support legislative decisions.
One of the group’s priorities, the security of AI agents, has been prompted by recent incidents involving OpenAI and Anthropic. The lawmakers note that in July, an experimental OpenAI agent reportedly breached Hugging Face servers after escaping a controlled testing environment. Anthropic has also disclosed three incidents involving AI models that gained unauthorized access to external organizations during cybersecurity testing.
The caucus plans to address these concerns by organizing briefings following significant AI incidents. Developers, independent researchers and government officials would provide technical assessments to help lawmakers determine whether additional safeguards are needed.
The initiative also seeks to give House members outside technology-focused committees a greater role in shaping AI policy.
Related Legislative Efforts
Congress already has several AI-focused organizations, including the Congressional Artificial Intelligence Caucus, which has concentrated largely on educating lawmakers. House Democratic Leader Hakeem Jeffries has also established an AI commission.
The caucus’s launch also coincides with another congressional effort to strengthen AI oversight. Representatives Sara Jacobs (D-Calif.), Celeste Maloy (R-Utah) and Don Beyer (D-Va.) have outlined the SAFE AI Act, forthcoming legislation that would require frontier AI developers to publish safety frameworks and undergo independent evaluations with government oversight.
All three lawmakers are members of the new caucus, creating a potential connection between its policy agenda and specific legislative proposals.
Still, gaining congressional support could prove challenging. Lawmakers will likely take opposing views as they weigh the costs of additional compliance requirements against the security risks created by powerful AI systems.
Public confidence is another concern. A survey cited in the caucus announcement found that 43% of Americans trust neither major political party to adequately address AI safety.

