Moran's AI Bill: Reporting Critical AI Incidents

U.S. Representative Nathaniel Moran has introduced new legislation called the Artificial Intelligence Incident Reporting Act. This bill would require reporting of critical AI incidents to the Secretary of Commerce. The proposed law aims to create a federal framework. This framework would require developers of complex AI models to report dangerous capabilities, security breaches, and safety incidents. Moran stated that AI needs accountability and human oversight. Under the bill, the Department of Commerce would determine which AI models meet specific capability thresholds. These thresholds would evaluate if a system's power poses significant risks to national security or public safety. The legislation also directs the Commerce Department to develop reporting thresholds in consultation with AI developers, academic researchers, and cybersecurity experts. Developers would need to file reports within seven days of discovering any dangerous activity. For critical reports, the Secretary of Commerce would notify congressional leadership and relevant committee chairs within 48 hours. These critical reports include instances like a model autonomously improving itself or attempts to evade human oversight. Brendan Steinhauser, CEO of The Alliance of Secure AI, emphasized the increasing risks from advanced AI models. He believes mechanisms are needed to identify these risks early. This legislation aims to ensure the U.S. government has information to act quickly when something goes wrong with high-capability AI systems.