For millions of Americans, the path to necessary medical care is often blocked by a bureaucratic hurdle known as prior authorization—a process requiring doctors to obtain insurer approval before a patient can receive a prescribed treatment, drug, or procedure. While intended to control costs and prevent unnecessary care, the system has become a source of profound frustration for physicians and patients alike. Now, artificial intelligence is being deployed to accelerate prior authorization decisions, but a growing chorus of voices—including the American Medical Association, the Electronic Frontier Foundation, and Stanford researchers—warns that AI may exacerbate wrongful denials and erode patient trust.

The Promise of AI: Speed and Efficiency

Proponents argue that AI can process vast amounts of data in seconds, automatically approving routine claims that meet clear criteria. According to a 2025 AMA survey, 61% of physicians worry that AI will increase denials of necessary treatments, but some health systems are already seeing benefits. For instance, Advocate Health has implemented AI to speed up prior authorizations, reducing administrative burden and cutting wait times. Major payers report that they have reduced prior authorization volumes by 11% through AI and other efficiencies, as noted by TechTarget.

In a notable development, an Independence Blue Cross team won first place at Google's Code for the Cure Hackathon with an AI prototype designed to make prior authorization clearer and fairer. The tool aims to provide transparent reasoning for decisions, potentially reducing disputes. However, such innovations remain isolated amid broader concerns.

The Peril: Bias, Denials, and Black Box Decisions

Critics argue that AI systems, often trained on historical claims data, can perpetuate existing biases and lead to higher denial rates for vulnerable populations. Stanford University researchers have raised alarms about the lack of human oversight in AI-driven insurance decisions, warning that automated systems may reject care without adequate justification. The AMA has highlighted that AI is already leading to more prior authorization denials, with physicians reporting that algorithms fail to account for individual patient nuances.

The Electronic Frontier Foundation (EFF) has taken legal action, suing the Centers for Medicare & Medicaid Services (CMS) over its AI prior authorization demonstration project. The EFF contends that the initiative violates patients' rights to due process by relying on opaque algorithms that cannot be meaningfully appealed. This lawsuit underscores a fundamental tension: while AI can handle straightforward cases, its black-box nature makes it difficult to challenge erroneous decisions.

Physician Frustration and Patient Harm

Physicians across the country are voicing concerns that AI is adding to their administrative burden rather than alleviating it. A report from Medical Economics notes that AI-driven systems often require additional documentation, increasing the time staff spend on prior authorization. The AMA's ongoing campaign to reform prior authorization has gained momentum, with doctors calling for standardized electronic processes and greater transparency.

Patients are the ultimate victims. Delays in care can lead to abandonment of treatment, worsened outcomes, and even death. The Undark article “Frustrated With Delays, Doctors Take Aim at Prior Authorization” captures the human toll: a woman with breast cancer waited weeks for approval of a critical scan, only to be denied by an AI system that flagged her case as low priority. Stories like these are driving legislative efforts, including proposed rules requiring insurers to disclose when AI is used in decision-making.

Regulatory and Legal Landscape

The debate is unfolding against a backdrop of policy changes. CMS is exploring expansion of prior authorization in Traditional Medicare, a move critics say is unsupported by evidence. Meanwhile, a bipartisan bill in Congress would mandate real-time prior authorization for Medicare Advantage plans and require insurers to explain denials. The EFF lawsuit adds pressure for federal oversight of AI in healthcare.

Experts emphasize the need for guardrails. “AI should augment, not replace, human judgment,” said Dr. Jesse Ehrenfeld, AMA president, in a statement. “Without transparency and accountability, these tools risk undermining the physician-patient relationship.”

Looking Ahead

As AI becomes more embedded in healthcare, the prior authorization controversy serves as a cautionary tale. The technology holds promise for reducing administrative waste, but only if implemented with rigorous testing, human oversight, and clear recourse for patients. The outcome of the EFF lawsuit and pending regulations will shape whether AI becomes a tool for efficiency or a source of new inequities.

For now, the answer to the question “Will AI fix prior authorization—or make it worse?” remains uncertain. What is clear is that the stakes could not be higher: the health and lives of millions depend on getting it right.