Consumer Protections for Artificial Intelligence Interactions
What changed between versions
The effective date for AI disclosure and compliance requirements was changed from January 1, 2027 to October 1, 2026, accelerating the implementation timeline by nearly a year.
New requirements were added for developers of high-risk AI systems to provide documentation, make public statements about their systems, and disclose algorithmic discrimination risks to the attorney general.
Developers must now make available documentation and information (like model cards or impact assessments) to deployers and third parties to help complete impact assessments.
A rebuttable presumption was added that developers used reasonable care if they complied with the new documentation and disclosure requirements in enforcement actions.
The bill now specifically targets high-risk artificial intelligence systems rather than all AI systems, focusing regulatory requirements on systems with greater potential for harm.