Artificial intelligence is already changing healthcare in ways most employees never see.
Behind the scenes, insurers and health plans increasingly rely on sophisticated software to sort information, identify patterns, automate administrative work and assist with decisions involving medical care. Used appropriately, those tools can help process enormous amounts of information faster and reduce some of the administrative friction that has long been part of the healthcare system.
But what happens when technology helps decide that a medical service should not be covered?
Georgia lawmakers have drawn an important line.
Beginning January 1, 2027, a new Georgia law will establish requirements governing how artificial intelligence and other automated tools may be used in certain utilization review decisions. The law does not prohibit insurers from using AI. Instead, it places a human checkpoint between an automated system and an adverse determination affecting a patient.
For Georgia employers sponsoring health benefits, the law is worth understanding even when the employer itself is not making utilization review decisions.
What Georgia’s New Law Actually Does
Senate Bill 444 addresses the use of artificial intelligence by private review agents and utilization review entities.
Under the new law, AI systems and other automated tools may still be used as part of the utilization review process. They can help automate tasks, reduce administrative burdens and participate in decision-making.
What they cannot do is independently deliver the final adverse determination.
Before an adverse determination is issued to a patient, the utilization review must be conducted by a qualified natural person, with a clinical peer participating in that review.
That distinction matters.
Georgia is not attempting to remove technology from healthcare decision-making. The law instead establishes a human-review requirement at the point where an automated process could result in care or coverage being denied.
The law takes effect January 1, 2027.
GEORGIA SB 444: THE HIGHLIGHTS
Effective date: January 1, 2027
AI is not prohibited. Insurers and utilization review organizations can continue using artificial intelligence and automated tools.
AI can assist with decisions. The technology may automate tasks, reduce administrative work and participate in the decision-making process.
An adverse determination requires human review. An AI system cannot issue the adverse determination until a qualified human reviewer conducts the utilization review.
A clinical peer must participate. The required human review includes participation by an appropriately qualified clinical peer.
Employers should ask questions now. Organizations preparing for 2027 renewals should understand how their carriers, TPAs and other health plan partners use automated decision-making in utilization review.
Why This Matters to Employers
Most employers do not sit at a desk approving or denying individual medical procedures. Those responsibilities typically reside with an insurer, third-party administrator or another organization involved in administering the health plan.
That does not mean employers should ignore how those decisions are made.
Employees rarely distinguish between the employer, insurance carrier, TPA, pharmacy benefit manager and the other organizations behind their benefits. When an employee has difficulty obtaining approval for care, the experience often becomes an issue for HR regardless of which organization actually made the decision.
That makes the growing use of AI in healthcare more than a technology story. It is also a benefits-management issue.
Employers already evaluate networks, plan design, costs, service capabilities and administrative performance when reviewing their health plans. Increasingly, another question belongs in that conversation: Where is automated decision-making being used, and what happens when the technology recommends denying care?
AI Can Be Useful Without Being the Final Word
There is an important balance here.
Healthcare administration involves extraordinary amounts of data. Automation can potentially identify information faster, flag cases for additional review and handle repetitive tasks that would otherwise consume significant staff time.
Employers should not assume that the presence of AI in a health plan is inherently a problem.
The concern is what authority the technology is given.
An algorithm may identify that a request does not appear to meet predetermined criteria. But medical care does not always fit neatly into predetermined categories. A patient’s medical history, previous treatment, complications and other circumstances may materially change what is appropriate in an individual case.
Georgia’s new requirement effectively preserves a human layer of judgment before an adverse determination reaches the patient.
That is a distinction employers may increasingly want to explore when evaluating vendors and plan partners, whether a particular arrangement is directly subject to SB 444 or not.
Questions to Ask Before the 2027 Plan Year
Employers do not need to become experts in artificial intelligence to have a useful conversation with their benefits partners.
A few practical questions can reveal quite a bit.
Ask your carrier, TPA or utilization review partner whether AI or automated decision-making tools are currently used in utilization review. If they are, determine what role those systems play and at what point a person becomes involved.
Employers may also want to ask:
- Can an automated system recommend an adverse determination?
- Who reviews that recommendation before it becomes final?
- What qualifications are required of the human reviewer?
- How are clinical peers incorporated into the process?
- How will procedures change when Georgia’s new requirements become effective?
- Are vendors updating contracts, workflows or documentation in preparation for 2027?
- What information is available to an employee who wants to understand or challenge an adverse determination?
These conversations can fit naturally into the broader review employers should already be conducting as part of their annual benefits consulting and renewal planning.
Fully Insured and Self-Funded Employers May Have Different Considerations
As with many insurance requirements, the application of a state law can depend on how an employer’s health plan is funded and regulated.
A fully insured health plan and a self-funded ERISA plan do not necessarily operate under the same state insurance requirements. Employers should therefore avoid assuming that SB 444 applies identically to every employer-sponsored health plan.
For self-funded organizations in particular, this is a good reason to involve benefits consultants, TPAs and legal or compliance resources when determining how a new state requirement interacts with the organization’s particular plan structure.
Even when a specific plan is not directly subject to a particular provision of state insurance law, the broader issue remains relevant. Employers can still ask how vendors use AI, how adverse determinations are reviewed and what safeguards exist before an employee receives a denial.
Technology Is Moving Faster Than the Traditional Benefits Conversation
There was a time when evaluating benefits technology primarily meant looking at enrollment systems, employee portals and electronic data feeds.
That definition is rapidly expanding.
Technology can now influence what employees see, how claims are processed, how potential fraud is identified, how utilization is evaluated and how healthcare decisions move through the system.
For employers, that creates an important distinction between using technology to make a process more efficient and allowing technology to make consequential decisions without meaningful human oversight.
Georgia’s new law puts that distinction into statute for covered utilization review activities.
It also provides employers with a timely reason to look more closely at technology that may otherwise remain largely invisible during the annual renewal process.
People Focused, Technology Empowered
MSI has long believed that technology is most valuable when it improves the experience of the people it is designed to serve.
That philosophy becomes particularly important when employees are dealing with a serious diagnosis, an expensive treatment or a disagreement over coverage. At those moments, efficiency matters, but so do context, communication and access to knowledgeable people.
MSI’s Member Advocacy team helps employees navigate coverage questions, claim disputes, provider issues and other benefit concerns while giving employers an experienced resource to help resolve problems that can otherwise become frustrating and time consuming.
As AI assumes a larger role in healthcare and insurance administration, employers will need to pay attention not simply to whether the technology is being used, but how it is being used, where human judgment remains part of the process and what happens when an employee needs help.
Georgia’s new law makes that conversation particularly timely.
Employers preparing for 2027 should use the months ahead to ask questions, understand their vendors’ processes and make sure that innovation and accountability continue moving forward together.
Want to better understand how changing regulations and emerging technology could affect your employee benefits program? Contact MSI Benefits Group to start the conversation.
External Resource: The Georgia General Assembly provides legislative information and documentation regarding Senate Bill 444 and other enacted legislation through its official website.