HB 474Just filedTechnology & privacy
Requires state agencies to keep human oversight of AI decisions and actions
Filed by
Salman BhojaniDDemocrat· House District 92· Filed Nov 9, 2026The short version · AI summary
State agencies that use artificial intelligence systems would have to set up policies so a trained employee or officer can review, approve, reject, modify, override or stop AI decisions. This covers decisions that are a controlling factor in a consequential decision, or that carry a foreseeable risk of substantial harm to health, safety, property, critical infrastructure or confidential information. The Department of Information Resources must adopt minimum standards by March 1, 2028. Agencies get 90 days after the rules to comply and must use interim safeguards until then. Effective September 1, 2027.
What it means for you
If a state agency uses AI to make an important decision about you or to run something risky, a trained person would have to be able to review and overrule it. Most people would not notice day to day.
One email when it moves, at most one a day. Unsubscribe in one click from any email.