Use of Artificial Intelligence in Health Care
What changed between versions
Added a new section prohibiting AI systems from making adverse coverage determinations (like denials of care) without meaningful review by a qualified human clinician or physician.
Mandated that AI systems used for utilization review must base determinations on individual medical history and clinical circumstances, explicitly forbidding reliance solely on group data.
Required entities to provide written disclosures to state agencies detailing how AI is used, the points in the process where it is applied, and the specific human oversight procedures in place.
Established new requirements for AI systems to produce and retain documentation, audit logs, and model-governance records to demonstrate compliance and ensure accuracy.
Added precise definitions for key terms such as 'behavioral health administrative services organization' and 'private utilization review organization' to clarify which entities must comply with the new rules.