Texas Congressman Proposes AI Incident Reporting Mandate

New federal reporting framework for AI risks

A Republican congressman has introduced legislation that would establish mandatory incident reporting for artificial intelligence companies, requiring them to disclose dangerous model behavior and security breaches to federal authorities.

Rep. Nathaniel Moran of Texas unveiled the AI Incident Reporting Act on Thursday, proposing a seven-day window for AI model developers to report critical incidents to the Commerce Department. The department would then have 48 hours to notify Congress of the most severe cases.

"It's a catch-it-early and sound-the-alarm bill," Moran said, according to Reuters, which first reported the legislation.

What would trigger reporting requirements

The draft legislation defines reportable incidents broadly. Companies would need to disclose when AI models attempt to evade human oversight, circumvent safety mechanisms, or otherwise undermine operator control. Unauthorized access to model weights—the parameters that guide an AI system's decision-making—would also require reporting.

The bill covers threats spanning chemical, biological, and nuclear risks to public safety, reflecting growing concerns about dual-use AI capabilities.

Why it matters

The proposal arrives at a moment when the federal government lacks a transparent framework for governing frontier AI systems. Earlier this month, the Commerce Department took action against Anthropic's latest models on national security grounds, forcing the company to disable global access. That incident exposed the absence of clear procedures for handling AI risks that cross the threshold from commercial concern to national security threat. A standardized reporting mechanism could provide both industry and regulators with earlier warning of emerging dangers.

Congressional gridlock on AI regulation

Congress has struggled to advance AI legislation amid competing priorities. Lawmakers remain divided over whether federal rules should override state laws and whether regulatory guardrails might slow American innovation in competition with China.

Earlier in June, two House members released a discussion draft of broader AI legislation called the Great American Artificial Intelligence Act, which also included provisions for reporting safety incidents to Commerce.

Moran believes his more focused approach could move faster through the legislative process and attract bipartisan backing. Mark Beall, president of the AI Policy Network, told Reuters that while "no legislation on AI has had much of a chance," he sees "growing demand from the public to see some action" and supports Moran's proposal.

The legislation represents one of several attempts to establish federal oversight of AI systems as their capabilities advance and their potential risks become more concrete.

Details of the proposed legislation were first reported by Reuters.