Bill > SB0205

This bill amends existing Utah law to establish new requirements for law enforcement agencies using artificial intelligence (AI), particularly generative AI, which is AI capable of creating content like text or images. Key provisions include defining "oversight or safeguard settings" for generative AI, which are features designed to ensure transparency, content safety, or human review, such as requiring human editing, disclosing AI authorship, inserting deliberate errors, or limiting AI use to specific crime types. Law enforcement agencies must now have a policy on generative AI use that is reviewed annually and includes details on which AI technologies can be used, their available safeguard settings, and the importance of human review of AI-generated content. Furthermore, any police report or law enforcement record created partly or wholly by generative AI must include a disclaimer and a certification from the author that they have reviewed it for accuracy. These policies must be publicly posted on the agency's website or the Utah Public Notice Website, with updates made within 30 days of any changes, unless the AI is used for internal prison or jail security, safety, or management. The bill also clarifies that failure to comply with these requirements does not automatically invalidate the use of the technology or the evidence it produces, and it sets an effective date of May 6, 2026.