S. 920 creates a new article in South Carolina insurance law governing how health insurers may use artificial intelligence, algorithms, or other software tools in prior authorization and utilization review decisions. The bill requires insurers to base such determinations on an enrollee’s medical history, the specific clinical circumstances identified by the treating provider, and other relevant medical-record information, rather than relying solely on group datasets or automated outputs. It also requires that any denial, reduction, or deferral of prior authorization be made by a licensed physician or other qualified healthcare professional who can evaluate the AI-generated recommendation in light of the enrollee’s individual clinical situation.
The bill adds disclosure and reporting obligations. Insurers must give prominent notice to enrollees that AI or similar tools are being used in utilization review, certify annually to the Department of Insurance that their systems are fair, accurate, reliable, and nondiscriminatory, and provide denial-rate information to the department, providers, and enrollees upon request. It also limits the use of patient data to the intended purpose consistent with HIPAA and authorizes the department to investigate alleged violations, hold hearings, require corrective plans, and impose existing disciplinary measures for repeat violations. The act would take effect upon gubernatorial approval.
The bill would amend Title 38 of the South Carolina Code by adding Article 25 to Chapter 71, directly regulating health insurers, health maintenance organizations, and the State Health Plan when they use artificial intelligence in prior authorization and coverage determinations. It would expand the Department of Insurance’s oversight authority by requiring annual certifications, disclosure practices, access to denial statistics, and compliance with standards against discrimination and improper data use. The measure would create new compliance duties for insurers and new enforcement tools for the department, while leaving the department authority to adopt implementing rules and regulations.
The available legislative history suggests generally favorable sentiment toward the bill. The Senate Banking and Insurance Committee reported the bill favorably and recommended that it pass, and the bill’s status notes only a scrivener’s error correction rather than any recorded opposition or divided vote. No committee transcript or roll-call vote is provided, so the record does not show detailed debate, but the committee action indicates support for regulating AI use in health insurance decision-making.
The main points of potential contention are the bill’s limits on insurer automation and the added compliance burden. Insurers may object to the requirement that adverse prior authorization decisions be made or reviewed by a licensed physician or other qualified healthcare professional, as well as the mandates for annual certification, disclosure, and production of denial-rate data. Supporters are likely to emphasize patient protections, individualized medical review, transparency, and anti-discrimination safeguards, while any critics would likely focus on administrative cost, operational complexity, and possible delays in utilization review.