This bill would require developers of advanced artificial intelligence (AI) models above certain thresholds to register with the federal government, adopt risk management, conduct independent assessments, publish redacted transparency reports, and maintain robust cybersecurity. Independent auditors would verify compliance, and federal authorities could issue emergency orders to restrict risky AI. State Attorneys General could help enforce the law, with civil penalties for violations. The bill also blocks states from passing new developer-focused rules in areas like risk management and incident reporting, but allows states to enforce general laws, regulate AI end users, protect minors, and manage their own AI use.
