US Lawmaker Proposes Bill for Mandatory AI Incident Reporting to Commerce Department

U.S. Representative Nathaniel Moran, a Republican from Texas, has put forward a significant piece of legislation aimed at enhancing oversight of artificial intelligence development. The proposed 'AI Incident Reporting Act' seeks to establish a mandatory reporting system for AI model developers, requiring them to disclose critical incidents to the U.S. Commerce Department. This move comes as Congress grapples with the complexities of regulating rapidly advancing AI technologies and their potential societal impacts. Under the terms of the draft bill, AI companies would be obligated to report any discovery of dangerous capabilities, security vulnerabilities, or safety incidents within seven days of their occurrence. Following such a report, the Commerce Department would then have a 48-hour window to notify Congress of the most serious incidents. Representative Moran emphasized that the bill is designed as a 'catch-it-early and sound-the-alarm' mechanism, highlighting the need for proactive measures to manage AI risks. This legislative effort is a direct response to growing concerns regarding the national security and public safety implications of increasingly sophisticated AI models. The absence of a transparent and standardized framework for governing frontier AI has become a pressing issue. An example cited is the Commerce Department's recent action against Anthropic's models, which led to their global disabling due to national security concerns, underscoring the need for clearer protocols. The bill defines reportable activities broadly, including instances where an AI model attempts to evade human oversight, circumvent established safeguards, or otherwise undermine human operators' control over the system. It also covers unauthorized access to model weights and threats to public safety involving chemical, biological, or nuclear elements. While Congress has seen various AI regulation proposals, Moran believes his targeted approach could garner bipartisan support and achieve a quicker path to becoming law. Mark Beall, president of the AI Policy Network, expressed support for the bill, noting a growing public demand for action on AI regulation.