AI in Health Insurance: Why Regulation Is Crucial
Artificial intelligence has the potential to improve care quality and reduce expenses in health insurance. However, it also carries risks like delays or outright denials of necessary care, often justified by cost savings. Over the past decade, insurers have increasingly relied on AI algorithms—not to treat patients, but to decide which recommended treatments get covered.
A common example of this is prior authorization. Before your doctor can provide certain care, they often need approval from your insurance company. Many insurers use AI to determine if the requested care is "medically necessary" and how much care a patient should receive, such as the number of hospital days after surgery. Unfortunately, evidence shows these systems can be used to delay or limit care that patients genuinely need.
When Claims Are Denied
If your insurer denies coverage for a recommended treatment, your options are limited:
- Appeal the decision, which can be costly, time-consuming, and require expert help. In reality, only 1 in 500 denials are appealed.
- Accept an alternative treatment that your insurer will cover.
- Pay out of pocket, which is often unrealistic due to high costs.
These hurdles mean many patients simply go without necessary care, raising serious health concerns.
A Pattern of Withholding Care
Insurers presumably input patient health records into AI systems and compare them against medical standards to decide coverage. But the algorithms remain a black box—insurers refuse to disclose how they work. This lack of transparency makes it impossible to assess their fairness or accuracy.
AI reduces the need for human review, saving insurers time and money. Beyond that, if a valid claim is denied and the patient appeals, the process can drag on for years. In severe cases, insurers might financially benefit if patients cannot wait and pass away before resolution.
Insurers often argue that patients can pay out of pocket if coverage is denied. This ignores reality, especially for expensive, long-term, or terminal illnesses. Older adults and people with chronic conditions face higher denial rates, and marginalized groups—including Black, Hispanic, and LGBTQ+ individuals—experience denials disproportionately.
Some research even suggests prior authorization can increase overall health system costs. Patients aren’t just being denied care—they’re being forced into impossible choices, with real consequences for their health.
The Current Regulatory Gap
Unlike AI tools used by doctors, insurance algorithms operate with little oversight. They don’t require FDA approval, and insurers claim their algorithms are trade secrets. This means there’s no public scrutiny or independent testing to verify safety, fairness, or effectiveness. There are no peer-reviewed studies to confirm these systems work well in real-life situations.
Some states—including Colorado, Georgia, Florida, Maine, and Texas—have proposed laws to regulate insurance AI. The Centers for Medicare and Medicaid Services (CMS) recently mandated that Medicare Advantage insurers base decisions on individual patient needs, not generic criteria. However, these rules still allow insurers to set their own standards and don’t require independent algorithm testing.
State laws often fall short by leaving too much control with insurers and not demanding outside review. Plus, states can’t regulate Medicare or insurers operating across state lines, limiting their impact.
The Case for FDA Oversight
There’s growing consensus among health law experts that AI used by insurers needs regulation. The FDA is well suited to step in because it already evaluates many medical AI tools for safety and effectiveness. FDA oversight would establish a consistent national standard rather than a confusing patchwork of state rules.
One challenge is that current FDA authority covers devices intended for diagnosis or treatment—not insurance decisions. Congress may need to update the legal definition to include coverage algorithms. Meanwhile, CMS and states could require independent safety, accuracy, and fairness testing to push insurers toward national standards.
Where Do We Go From Here?
Regulating AI in health insurance is no longer optional—it’s essential to protect patients. Without transparency and oversight, AI risks becoming a tool to prioritize profits over care. Stronger regulation can help ensure that AI supports fair, efficient coverage decisions, rather than denying necessary treatment.
For professionals in insurance, staying informed about these developments is key. Understanding how AI impacts coverage decisions can help you advocate for better policies and serve clients more effectively.
Learn more about AI and its applications in insurance at Complete AI Training.
Your membership also unlocks:
