The Georgia Senate has unanimously passed legislation aimed at preventing health insurance companies from using artificial intelligence (AI) as the sole basis for denying medical coverage, requiring instead that a human clinician make the final decision.
The bill, Senate Bill 444, mandates that “decisions with regard to the provision of insurance coverage for healthcare services shall not be based solely on artificial intelligence systems.” While insurers may still utilize AI to automate administrative tasks and assist in the review process, the legislation stipulates that technology cannot override the judgment of a clinician.
State Senator Kay Kirkpatrick, a retired orthopedic surgeon and the bill’s sponsor, emphasized the necessity of human oversight in the medical claims process. “There needs to be a human review of the information before they turn you down for your treatment that your doctor ordered,” Kirkpatrick said.
The legislative push comes amid growing concerns over the automation of healthcare denials. A survey by the National Association of Insurance Commissioners found that 43% of 93 large health insurers already use or plan to use AI for claim adjudication, while 26% intend to use it for benefit determinations.
Consumer advocates have signaled that the stakes are high for patients. Liz Coyle, Executive Director of Georgia Watch, noted that 20% of all claims across Affordable Care Act marketplace plans were denied in 2023. Coyle argued that allowing a “robot” to make life-and-death decisions regarding potentially life-saving procedures is “completely unacceptable.”
However, some lawmakers expressed concern that the current language might allow for superficial oversight. State Representative Todd Jones (R-Forsyth) cautioned that a “clinical peer could, for all intents and purposes, review it for a microsecond” and sign off on an AI decision without performing a substantive review.
The bill defines AI as a “machine-based system” that uses human-defined objectives to make predictions, recommendations, or decisions. While Kirkpatrick noted that doctors are generally optimistic about AI’s potential to streamline insurance bureaucracy, she maintained that the technology is not yet ready to replace human medical professionals.
The legislation is scheduled for a vote before the House Technology committee on March 17. If signed into law, the new regulations are slated to take effect in 2027.