US House Passes Bipartisan AI Governance Package 312-115

The U.S. House of Representatives passed a bipartisan AI governance package on May 7, a bundle of four bills establishing minimum transparency requirements for AI systems in federal use, export control guidelines for frontier models, and a new AI Safety Commission with subpoena authority.

Key points:

• The package passed 312-115, with majority support from both parties, making it the most significant federal AI legislation since the executive orders of 2023-24.

• The AI Safety Commission would have authority to investigate AI incidents, issue safety recommendations, and compel cooperation from AI developers in federally related deployments.

• Export control provisions add statutory authority to existing Commerce Department guidance, making frontier model export restrictions harder to reverse through executive action alone.

Bipartisan passage with a 312-115 margin is a signal that AI governance has crossed from partisan debate to policy consensus in Congress. The legislative momentum will now be harder to slow. The subpoena authority granted to the AI Safety Commission is the most consequential provision: it means AI incidents involving federal systems can now be formally investigated, not just self-reported.

For AI governance leaders, the Safety Commission's framework is a preview of what internal governance processes will need to match. Begin building incident response documentation that would satisfy a formal investigation. Track the package through Senate debate: the Senate Commerce Committee will likely amend it significantly. The final bill is the one that matters for compliance planning.

Why It Matters: Bipartisan 312-115 passage signals AI governance has crossed from partisan debate to policy consensus, with the Safety Commission's subpoena authority enabling formal investigations of AI incidents rather than relying on self-reporting.