The Georgia General Assembly adjourned on April 3, 2026, sending three AI-related bills to Governor Brian Kemp’s desk. None have been signed yet, but they represent some of the most substantive state-level AI legislation to emerge so far this year.
The headline is SB 444, which would ban health insurers from making coverage denial decisions using AI alone. It is a practical, targeted bill that has flown under the radar nationally but has real implications for any business providing employee health coverage.
SB 444: AI Cannot Be the Final Word on Health Insurance Denials
The bill amends Georgia’s laws on “private review agents” — the entities that conduct utilization review for health insurers. Under SB 444, any adverse coverage decision must involve a qualified human clinician before it is finalized. AI tools can still be used to speed approvals, detect fraud, process paperwork, and surface relevant clinical criteria. But an AI system or software tool cannot be the sole basis for denying a claim.
The sponsor, Sen. Kay Kirkpatrick (a retired orthopedic surgeon), has been direct about why: AI systems are good at pattern-matching against historical data but are not equipped to weigh the full clinical picture of an individual patient. Denials have downstream effects on people’s health, and those decisions need a human being accountable for them.
The timing reflects where the industry is heading. A National Association of Insurance Commissioners survey found that 43% of large health insurers already use or plan to use AI for claim adjudication, and 26% for benefit determinations. SB 444 draws a clear line before that trend runs ahead of accountability.
SB 540: Chatbots Must Say They Are Chatbots
The second major bill requires AI chatbots to disclose they are not human — every three hours for adult users, every hour for accounts belonging to minors. Operators must also provide parental controls for underage users, publish crisis referral protocols, and implement safeguards around content involving suicidal ideation or self-harm.
The bill is deliberately scheduled to take effect in 2027, giving Congress time to act first. Georgia’s legislature has been explicit about not wanting to get ahead of federal standards, but it is not waiting forever.
SR 789: An AI Study Committee
The third measure is a Senate resolution creating a study committee to examine AI’s impact on Georgia. It builds on a 2024 committee that produced a 185-page report with 22 recommendations. This is not enforceable law, but it suggests more substantive legislation is being drafted for future sessions.
What This Means for Business
If you employ people in Georgia and provide health insurance, SB 444 matters to you. Not because it changes what your employees can claim, but because it establishes that whoever is administering your plan cannot outsource coverage decisions entirely to an algorithm. That is a reasonable expectation, and similar bills are moving in other states.
More broadly, this is part of a pattern worth paying attention to. State legislatures are not trying to stop AI. They are drawing lines around specific high-stakes decisions where accountability has to rest with a human. Insurance denials are one. Mental health chatbots targeting minors are another. The legislation is sectoral, not sweeping.
For businesses building or deploying AI, this is the direction regulation is heading: not blanket prohibitions but specific rules about where human oversight is non-negotiable. Building your AI systems with that in mind from the start is far cheaper than retrofitting compliance later.
Georgia joins a growing list of states taking targeted action. Tennessee signed a law in April barring AI from posing as licensed mental health therapists. Idaho passed the Conversational AI Safety Act for consumer-facing chatbots. The federal landscape remains contested, but the states are not waiting.
Enterprise DNA put together a free field guide on exactly this: the full Claude ecosystem, Claude Code, and how to roll agents out without breaking things. Get the guide.