AI Incident Reporting Act

Bill Number

H.R. 9477

Origin Chamber

House

Congress

119th Congress, Session 2

Policy Area

Science, Technology, Communications

Status

Introduced

Latest Action

2026-06-25: Referred to the House Committee on Energy and Commerce.

Last Updated

2026-07-08T18:56:39Z

AI-Generated Summary

AI Incident Reporting Act (H.R. 9477)

Purpose

This legislation establishes a mandatory reporting system for certain high-risk artificial intelligence models. It aims to ensure that developers notify the federal government about incidents or capabilities that could threaten U.S. national security or public safety.

Key Provisions

• Designation Process: Within 180 days of enactment, the Secretary of Commerce must issue regulations defining thresholds for "covered models" and "covered model developers" based on risks to national security or public safety. This involves consultation with agencies, experts, and industry.
• Reportable Activities: Developers must report specific events within 7 days, including:
• Models attempting to evade human control, deceive operators, or resist shutdown.
• Unauthorized access, theft, or exfiltration of model weights (the internal parameters needed to run the model).
• Capabilities enabling large-scale cyber attacks on critical infrastructure.
• Autonomous acceleration of advanced AI development.
• Assistance in developing chemical, biological, radiological, nuclear, or explosive weapons.
• Near-misses where serious harm was narrowly avoided.
• Reporting Requirements: Initial reports must detail the incident, discovery date, potential causes, and national security implications. Supplemental reports are required as new information emerges. Expedited reporting applies to imminent risks.
• Congressional Notification: The Secretary must inform congressional leaders and relevant committee chairs within 48 hours for imminent risks or 30 days for other reports.
• Information Protections: Submitted data is exempt from public disclosure laws, cannot be used against the developer in most legal actions, and may be shared within government with safeguards.
• Enforcement: The Secretary can investigate, issue subpoenas, and impose civil penalties up to $2 million per violation, with each day of noncompliance counting separately.

Significant Changes to Existing Law

This bill introduces new federal oversight requirements for AI development, as no prior law mandates such incident reporting for artificial intelligence models. It creates exemptions from disclosure under the Freedom of Information Act and limits how government agencies can use the reports in enforcement actions, while preserving liability for underlying incidents based on independent evidence.

Potential Impacts

• Government Agencies: Expands the Department of Commerce's role in AI oversight, requiring coordination with intelligence and law enforcement entities for analysis and response.
• Citizens and Public Safety: Aims to reduce risks from advanced AI by enabling early government intervention, though it does not directly affect everyday users.
• International Relations: Could influence global AI standards if U.S. requirements set precedents, but the bill focuses solely on domestic reporting without addressing foreign entities.

Main Stakeholders Affected

• Covered AI model developers (those training or significantly modifying high-capability models).
• The Secretary of Commerce and relevant federal agencies.
• Members of Congress and congressional committees.
• Experts in AI, cybersecurity, and national security consulted during rulemaking.

Notable Legal, Constitutional, or Political Implications

The bill includes strong protections against using reports in civil or criminal proceedings to encourage compliance, but it raises questions about balancing government access to sensitive AI information with developer confidentiality. It grants broad enforcement authority to the executive branch without creating new criminal penalties, and the rulemaking process may involve debates over defining risk thresholds.

This summary was generated by AI and may contain inaccuracies. Refer to the official source document for the authoritative text.

Sponsor

Rep. Moran, Nathaniel [R-TX-1]

Recent Actions

• 2026-06-25: Referred to the House Committee on Energy and Commerce.
• 2026-06-25: Introduced in House
• 2026-06-25: Introduced in House

Bill Versions

• AI Incident Reporting Act — issued 2026-06-25 — PDF (16 pages)