MN SF4509

Requires AI developers to implement written safety and security protocols before deploying AI models, including cybersecurity protections, testing procedures, and designation of senior compliance personnel

Defines "critical harm" as death, serious injury, or mental injury to 25+ people, or at least $1,000,000 in property damages, caused by AI involvement in weapons creation or autonomous criminal conduct

Mandates developers report safety incidents to the attorney general within 72 hours and retain safety protocol records for the deployment period plus five years

Prohibits deployment of AI models that create unreasonable risk of critical harm and bars developers from making false or misleading statements in required documentation

Authorizes civil penalties up to $10,000,000 for first violations and $30,000,000 for subsequent violations, plus establishes a private right of action for injured persons to recover damages and attorney fees