The Digital Gatekeeper: Will Artificial Intelligence Streamline Healthcare Prior Authorization or Compound Existing Barriers?

The American healthcare system is currently at a pivotal crossroads as the federal government and private insurers increasingly turn to artificial intelligence (AI) to manage one of the most contentious aspects of medical administration: prior authorization. This process, which requires healthcare providers to obtain advance approval from a patient’s insurance plan before specific treatments, medications, or procedures are administered, has long been a source of friction between medical professionals and payers. While the Trump administration and industry leaders argue that AI can modernize this "broken" system by expediting approvals and reducing waste, a growing chorus of physicians, patient advocates, and lawmakers warns that automating these decisions could lead to a surge in wrongful denials and further jeopardize patient health.

The Evolution and Burden of Prior Authorization

Prior authorization (PA) was originally conceived as a clinical safeguard to ensure that medical services are necessary, safe, and cost-effective. By requiring a secondary check on high-cost treatments or those with lower-cost alternatives, insurers aimed to curb unnecessary medical spending and prevent the overuse of potentially harmful interventions. However, over the past decade, the scope of PA has expanded significantly, now covering everything from routine diagnostic imaging to life-saving oncology medications.

The administrative burden associated with this process is substantial. According to a 2025 survey conducted by the American Medical Association (AMA), the vast majority of physicians report that PA requirements have a significant negative impact on clinical outcomes. The survey revealed that 61 percent of doctors are concerned that the integration of AI will only exacerbate the rate of denials for treatments they deem medically necessary. For patients, the consequences of these delays can be dire. Data from the Commonwealth Fund indicates that in 2025, approximately one in five working-age adults with private insurance reported that they or a family member were denied coverage for physician-recommended care. Of those denied, 41 percent experienced a delay in care, and more than a quarter reported that their underlying health condition worsened as a direct result of the wait.

The Promise of Artificial Intelligence in Claims Processing

Proponents of AI integration argue that the technology is uniquely suited to solve the "prior auth purgatory" that currently plagues the system. By using machine learning algorithms to scan vast quantities of medical records and clinical guidelines, AI can theoretically identify "unambiguously allowable" claims in seconds—a process that currently takes human reviewers days or even weeks.

The theoretical goal is to create a "gold-carding" system where physicians with high approval rates or routine, evidence-based requests receive instantaneous authorization. This would allow human reviewers to focus their attention on complex cases that require nuanced clinical judgment. Camm Epstein, a prominent health policy analyst, noted that "AI should be used to make appropriate care easier to approve, not necessary care easier to deny." If implemented correctly, AI could eliminate the manual faxing and phone calls that still characterize much of medical billing, reducing administrative overhead for both hospitals and insurance companies.

The WISeR Model: A Federal Pilot Program

The Trump administration has moved beyond theory by launching a significant pilot program titled the Wasteful and Inappropriate Service Reduction (WISeR) Model. Operated through the Centers for Medicare and Medicaid Services (CMS), WISeR is a demonstration project designed to leverage AI and machine learning to target waste, fraud, and abuse within the original Medicare framework.

Running through December 2031, the project is being piloted in six states and focuses on services identified as being highly vulnerable to overuse. These include:

• Skin and tissue substitutes
• Electrical nerve stimulator implants
• Knee arthroscopy for knee osteoarthritis

The WISeR model represents a significant shift because prior authorization has historically been a hallmark of private Medicare Advantage plans, rather than original Medicare. By combining automated algorithms with human clinical oversight, CMS aims to ensure that Medicare payments are "timely and appropriate." However, the program has already faced significant political and professional pushback. Critics argue that by introducing PA into original Medicare, the government is erecting new barriers for a population—seniors and the disabled—that previously enjoyed relatively unfettered access to care.

A Chronology of Reform and Policy Shifts

The push to reform prior authorization has seen a flurry of activity across two administrations. The timeline reflects a growing recognition that the current system is unsustainable:

• January 2024: The Biden administration issued a landmark rule requiring government-run health plans to streamline PA processes. This included a mandate that insurers respond to urgent requests within 72 hours and non-urgent requests within seven calendar days.
• January 2025: These timeline requirements officially went into effect for most public sector health plans, marking a major shift toward faster accountability.
• Early 2025: The Trump administration, in collaboration with private insurers, secured an industry pledge to further accelerate PA processes. Private companies vowed to standardize electronic requests by 2027 and reduce the volume of services subject to PA—such as colonoscopies and cataract surgeries—by 2026.
• June 2025 to April 2026: An industry-based survey reported an 11 percent decline in the total volume of prior authorization requests, suggesting that the pressure from the executive branch may be yielding results.

Despite these efforts, the "two minds" of federal policy remain a point of confusion. While CMS Administrator Mehmet Oz has warned private insurers to "ease the burden" of PA or face strict federal regulation, the same agency is simultaneously expanding PA within original Medicare through the WISeR AI pilot.

Statistical Realities and the Denial Crisis

While the volume of PA requests may be trending downward, the rate and impact of denials remain a primary concern. In Medicare Advantage—which now enrolls roughly 55 percent of Medicare-eligible individuals—insurers issue millions of full or partial denials annually.

A 2022 memorandum from the Department of Health and Human Services (HHS) Office of Inspector General (OIG) found that in more than 10 percent of cases, Medicare Advantage plans denied access to services that actually met Medicare coverage rules. This suggests that the "gatekeeping" function of PA often overreaches, blocking medically necessary care. Interestingly, the data also shows a high success rate for those who fight back: in 2024, Medicare Advantage plans overturned 81 percent of their own denials upon appeal. However, the complexity of the appeals process means that many patients simply give up, leading to what some call "rationing by bureaucracy."

The Ethical Dilemma: Profits vs. Patient Care

One of the most controversial aspects of the WISeR model involves the financial incentives provided to the vendors hired to run the AI systems. These vendors earn a share of what CMS classifies as "averted expenditures." This revenue model has drawn sharp criticism from health reform advocates like Wendell Potter and researchers like Zena Wolf.

Critics contend that paying vendors based on how much money they "save" the government creates an inherent conflict of interest. If a vendor’s profit is tied to rejecting care requests, the incentive is to tune AI algorithms toward denial rather than approval. This has led to accusations that the government is essentially "outsourcing" the denial of care to black-box algorithms that lack transparency. Several lawmakers have already introduced resolutions and amendments to block funding for the WISeR model, citing it as a threat to patient access and a violation of the physician-patient relationship.

Analysis of Broader Implications and the "Arms Race"

As AI becomes more deeply embedded in the healthcare infrastructure, experts warn of a looming "arms race" between payers and providers. Jared Dashevsky, a physician and founder of Healthcare Huddle, suggests that instead of using AI to fix the underlying issues, both sides are simply automating their existing combat. Insurers are using AI to deny claims faster, while hospitals are beginning to use AI to generate appeals faster.

"This is more automation of a broken system that shouldn’t exist in its current form," Dashevsky noted. If the future of healthcare involves two sets of algorithms arguing with each other while the patient waits for treatment, the administrative "savings" promised by AI may never materialize. Instead, the system may simply become more opaque, making it harder for patients and doctors to understand the clinical reasoning behind coverage decisions.

Furthermore, the lack of transparency in AI algorithms remains a major hurdle. The AMA has advocated for a "human-in-the-loop" requirement, insisting that no denial of care based on medical necessity should be finalized without a review by a qualified physician. While insurers have recently pledged that AI is not used to deny requests without clinical review, the sheer volume of claims and the speed of AI processing raise questions about how thorough those human reviews can truly be.

Conclusion

The integration of AI into the prior authorization process represents a high-stakes gamble for the American healthcare system. On one hand, it offers a path toward a 21st-century administrative model that replaces faxes and delays with instant, data-driven decisions. On the other, it risks codifying a "culture of denial" into the very code of the healthcare system, where financial incentives for "averted expenditures" outweigh clinical necessity. As the WISeR pilot continues through 2031, the results will likely determine whether AI becomes a tool for healthcare efficiency or a digital wall between patients and the care they need. For now, the system remains in a state of uneasy transition, with the federal government simultaneously acting as both a reformer of private industry and a pioneer of automated gatekeeping in the public sector.