Georgia lawmakers are drawing a line on robot-only health insurance denials, advancing a bill that would require a human clinician to sign off before coverage can be turned down. The state Senate passed SB 444 unanimously, and the measure is now headed to the House Technology Committee next Tuesday. Supporters say the bill would keep a doctor involved in tough calls while still letting insurers lean on artificial intelligence for everyday administrative work.
What SB 444 Would Change
SB 444 would amend Chapter 46 of Georgia law to explicitly bar coverage determinations that rely solely on artificial intelligence outputs. Before any adverse coverage decision is finalized, a licensed clinical peer would have to be involved, according to the bill text. Sponsor Sen. Kay Kirkpatrick (R-Marietta), a retired orthopedic surgeon, has said the aim is to shield patients from automated denials while still allowing insurers to use AI behind the scenes for tasks like paperwork.
The legislation moved through the Senate unanimously and then crossed over to the House for further consideration, as outlined by the Georgia General Assembly.
How Insurers Already Use AI
A survey of 93 large health insurers by the National Association of Insurance Commissioners found that about 84% reported some use of AI or machine learning in their operations, and roughly 43% said they already use or plan to use it for claims adjudication. About a quarter reported using or exploring AI tools for benefit-eligibility checks, according to the NAIC. The study details deployments across prior authorization, fraud detection and claims routing, and notes that many companies rely on third-party AI components.
Meanwhile, data compiled by KFF shows that nearly one in five in-network claims were denied in 2023, a statistic advocates frequently cite as they warn about the potential harm of automating more of the denial process.
Lawmakers, Doctors and Advocates Push Back
Kirkpatrick told GPB she wants “a human clinical peer in the process before an adverse determination is made,” arguing that patients should not be rejected by software alone. Consumer advocates have echoed that concern. Georgia Watch Executive Director Liz Coyle warned that AI could make already-high denial rates worse, and some House committee members pressed on whether a clinical peer might just rubber-stamp denials instead of providing real scrutiny, as reported by WSB-TV.
Those exchanges in the Senate and as the bill moves through House hearings have highlighted a broader anxiety about how far automated decision-making should go in health care.
Why Supporters Say It Will Not Smother Helpful AI
Supporters of SB 444 stress that the proposal does not outlaw AI in health insurance. They argue insurers could still use algorithms to speed approvals, flag possible fraud or cut down on paperwork. The line the bill draws is that AI cannot be the first and last word on coverage decisions that can shape a patient’s treatment.
Researchers at Stanford have warned that insurers are rapidly adopting AI for utilization review while governance and oversight lag, a concern lawmakers have pointed to in crafting SB 444. The National Association of Insurance Commissioners has been examining how insurers deploy these tools and has published materials to guide regulators on when human review should be required, according to the NAIC.
What’s Next
The House Technology Committee is scheduled to take up the bill next Tuesday, according to WSB-TV. If SB 444 clears that panel, it would head to the full House for a vote.
Backers argue the measure strikes a workable balance between oversight and efficiency. Critics want tighter definitions and stronger enforcement language to make sure human review is more than a cursory sign-off. Patients, providers and policy groups are watching the upcoming committee debate closely for how the state will ultimately define and audit what counts as “human review.”
