HB 306 Summary

• Requires persons who distribute deepfakes of candidates or political parties in campaign communications to include a disclosure statement: "This communication has been manipulated or generated by artificial intelligence," with specific formatting requirements for audio, print, and video formats.

• Mandates state agencies conduct biennial inventories of artificial intelligence systems used for consequential decisions and publish them on department websites, including system names, vendors, capabilities, and impact assessment completion status.

• Requires state agencies to conduct impact assessments of AI systems at least every two years, analyzing efficacy, human oversight, accountability mechanisms, appeals processes, risks, effects on individual liberty and privacy, and potential discrimination.

• Prohibits state agencies from using AI systems for consequential decisions that involve biometric identification, emotion recognition, cognitive behavioral manipulation, or social scoring, and restricts use of systems with data hosted in China, Cuba, Iran, North Korea, Russia, or Venezuela.

• Establishes civil liability for individuals harmed by violations of AI requirements, allowing recovery of damages, punitive damages, and attorney fees in superior court against the state or state employees.