The Prior Authorization AI Arms Race: What It Means for Your Imaging Orders
Your doctor orders an MRI. Before the scanner can be scheduled, a request goes to your insurer. Somewhere in that insurer’s system, an algorithm reviews the request against coverage criteria. In a growing number of cases, that algorithm — not a physician — decides whether your imaging study is approved or denied.
This is not a future scenario. It is the present reality of prior authorization in American healthcare. And in 2026, it is accelerating.
The AI takeover of prior authorization
Prior authorization — the requirement that physicians get insurer approval before ordering certain tests, procedures, or medications — has been a source of friction in healthcare for decades. What has changed is who, or what, is making those decisions.
Thirty-seven percent of insurers now report using AI for prior authorization, with that number climbing rapidly. In the employer group market, the figure is even higher: 70 percent are using or exploring AI for prior authorization decisions. AI spending on prior authorization grew tenfold in a single year, from $10 million in 2024 to $100 million in 2025.
The scale is staggering. Large private insurers — including UnitedHealthcare and Humana — are deploying algorithmic systems to process millions of authorization requests. These systems can review a request in seconds, cross-reference it against clinical guidelines, and return a decision before a human reviewer has finished reading the first page.
The promise is efficiency. The concern is that efficiency cuts both ways.
Medicare joins the race: the WISeR pilot
On January 1, 2026, the Centers for Medicare and Medicaid Services launched the WISeR (Wasteful and Inappropriate Service Reduction) model — a six-year pilot program that introduces prior authorization to Original Medicare for the first time in selected regions.
The program operates in six states: New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington. Contractors use artificial intelligence, machine learning, and algorithmic decision logic to screen authorization requests. For approved requests, AI can handle the decision end-to-end. For denials, the model requires a human clinician with relevant expertise to conduct the medical review — an important safeguard, but one that does not eliminate the AI’s role as the first gatekeeper.
The initial services targeted are specific procedures — spinal stimulator implants, epidural steroid injections, vertebral augmentation, and percutaneous lumbar decompression. These are not imaging services. But the infrastructure being built is designed to expand. If the pilot demonstrates cost savings, the model’s service list will grow. And advanced imaging — MRI, CT, PET — has historically been one of the most heavily prior-authorized categories in commercial insurance.
For physicians and practice managers who have watched prior authorization expand relentlessly over the past two decades, the WISeR pilot represents a significant shift: the federal government is now building the same AI-driven authorization machinery that commercial insurers have been deploying for years.
Why physicians are alarmed
The American Medical Association has been tracking the impact of prior authorization for years. The data paints a consistent picture, and it is getting worse.
An AMA survey released in 2025 found that 61 percent of physicians believe insurers’ use of AI is increasing prior authorization denial rates. This is not a theoretical concern. Physicians are seeing it in their daily workflows:
- 93 percent of physicians report that prior authorization delays necessary care.
- 82 percent say patients have abandoned recommended treatment due to prior authorization burdens.
- Physicians and their staff complete an average of 39 prior authorization requests per week, consuming roughly 13 hours of staff time per physician.
- 89 percent of physicians say the prior authorization burden contributes to burnout.
The imaging-specific numbers are particularly troubling. Radiology services face some of the highest denial rates in outpatient care, with up to 28 percent of claims initially denied. For a referring physician ordering an MRI to evaluate a suspected disc herniation or a CT to rule out a pulmonary embolism, a denial does not just create paperwork. It creates a gap in care.
What happens when imaging gets delayed
Prior authorization delays are not administrative inconveniences. They are clinical events with measurable consequences.
Research from Johns Hopkins Medicine documented measurable patient harm linked to prior authorization delays. When the interval between ordering a diagnostic study and completing it increases, conditions worsen. Patients who needed urgent imaging end up in emergency departments. Surgical planning stalls. Cancer staging gets delayed.
The AMA’s own data quantifies the downstream effects:
- 23 percent of physicians report that prior authorization has led to a patient hospitalization.
- 18 percent say it has caused a life-threatening event.
- Seven in ten physicians say their patients face delays of a week or longer when prior authorization is required.
- 87 percent report that prior authorization leads to higher overall healthcare utilization — the exact opposite of its stated purpose.
The irony is hard to miss. A system designed to reduce unnecessary utilization is generating emergency visits, repeat appointments, hospitalizations, and redundant testing. The administrative cost of prior authorization is estimated to exceed $35 billion annually across the US healthcare system. The human cost is harder to quantify but no less real.
The arms race dynamic
Here is where it gets interesting — and concerning. As insurers deploy AI to deny claims faster, providers are beginning to deploy AI to fight denials faster. Health Affairs has described this as an “AI arms race” in utilization review, and the metaphor is apt.
On one side, payer AI systems screen authorization requests against clinical criteria, flagging those that do not meet algorithmic thresholds. These systems can process thousands of requests per hour, applying rules that may or may not reflect the nuance of individual patient situations.
On the other side, provider organizations are investing in AI tools that optimize authorization submissions — pre-checking clinical documentation against known payer criteria, auto-generating appeal letters, and predicting which requests are likely to be denied so they can be preemptively strengthened.
The result is an escalating cycle of algorithmic optimization on both sides, with patients caught in the middle. Neither side’s AI is making the patient better. Both sides’ AI is consuming resources that could be spent on care.
The AMA has called for federal regulation of algorithm-based utilization review. Resolution 226 specifically seeks oversight to “transparently evaluate potential biases in how AI systems review and deny requests for diagnostic tests, procedures, and medications.” Beginning in 2026, CMS will require payers to provide a specific reason for every AI-assisted denial and to publish aggregate approval data — a step toward accountability, but not yet a solution.
What this means for your practice
If you are a referring physician or practice manager, the prior authorization landscape in 2026 demands a different approach than it did even two years ago.
1. Document aggressively. AI-driven prior authorization systems evaluate requests against structured criteria. Incomplete or vague clinical documentation is the single most common reason for algorithmic denial. Ensure every imaging order includes the specific clinical indication, relevant history, prior treatments attempted, and the clinical question the study is intended to answer.
2. Know the criteria. Each payer’s AI is trained on that payer’s coverage policies. Familiarize yourself with the clinical decision support criteria used by your top payers. Many follow ACR Appropriateness Criteria or similar evidence-based guidelines. If your order aligns with published criteria, say so explicitly in the request.
3. Appeal strategically. One in five physicians always appeals adverse prior authorization decisions. Appeals succeed at surprisingly high rates — often 50 to 75 percent for imaging denials. The initial denial is frequently an algorithmic threshold, not a final clinical judgment. Build appeal workflows into your practice operations rather than treating them as exceptions.
4. Speed matters for appeals. When you are appealing a denied imaging order, every day counts. Having the clinical evidence — including any prior imaging studies — immediately accessible and shareable accelerates the process. If you are waiting for CDs to arrive from another facility, or trying to get a patient’s prior scans released from a portal, you are losing time that affects the appeal window and the patient’s care.
5. Track your data. Monitor your practice’s prior authorization denial rates by payer and by service. If a specific payer is denying a disproportionate share of imaging requests, that pattern is worth raising with your payer representative — and potentially with your state medical association.
How faster results and instant sharing change the equation
The prior authorization arms race creates a specific problem for imaging workflows: delays compound. A denied authorization delays the scan. A delayed scan delays the diagnosis. A delayed diagnosis delays the treatment. And when an appeal requires documentation from prior studies at other facilities, the delay compounds further.
This is where the combination of AI-assisted analysis and instant digital sharing becomes clinically relevant — not as a solution to the prior authorization system itself, but as a way to minimize the damage it causes.
When imaging results are available in under three seconds through AI-powered analysis, physicians have the clinical data they need to support urgent authorization requests and appeals. A preliminary AI read that flags a suspicious finding provides documented clinical urgency that strengthens the case for expedited review.
When prior studies can be shared instantly through platforms like Medixshare by AI Bharata — without waiting for CDs, faxes, or portal access — appeal timelines shrink. A physician appealing a denied MRI who can immediately share the patient’s prior CT from another facility has a stronger, faster case than one who is waiting three days for records to arrive.
The friction of medical scan sharing has always been a problem. In the context of prior authorization appeals, where time directly affects patient outcomes, that friction becomes a clinical risk factor.
What patients can do
As a patient, you have more power in the prior authorization process than you might think.
Ask about prior authorization upfront. When your doctor orders imaging, ask whether it requires prior authorization and how long approval typically takes. If your condition is urgent, ask your physician to request an expedited review.
Request your records proactively. If you have had imaging done at another facility, do not wait for your new doctor to request the records. Share your scans directly so they are available immediately when needed for authorization requests or appeals.
Know your appeal rights. If an imaging order is denied, you have the right to appeal. Ask your physician’s office to file the appeal, and request a copy of the denial letter. Under the new 2026 CMS requirements, payers must provide a specific reason for AI-assisted denials.
Do not abandon care. Eighty-two percent of physicians say patients have abandoned recommended treatment due to prior authorization burdens. If your imaging order is denied, it does not mean the test is unnecessary. It means an algorithm flagged it for review. Work with your physician’s office on the appeal.
The road ahead
More than 50 major insurers have pledged to streamline prior authorization starting in 2026, with a goal of processing at least 80 percent of electronic prior authorization approvals in real time. Whether those pledges translate into meaningful change remains to be seen.
The CMS WISeR pilot will run through 2031, providing years of data on how AI-driven prior authorization affects utilization, costs, and patient outcomes in Medicare. If it expands to include imaging services, radiology will be at the center of the debate.
What is clear is that prior authorization is not going away, and AI is making it faster — for better and for worse. The question is whether that speed will be used to approve appropriate care more quickly or to deny it more efficiently.
For physicians, the practical response is to document thoroughly, appeal consistently, and ensure that clinical data — including imaging studies — moves as fast as the algorithms that evaluate it.
For patients, the response is simpler: do not let a denied authorization become a delayed diagnosis. Your scans are yours. Share them instantly using tools built for exactly this purpose — AI Bharata’s Medixshare makes it possible in seconds — keep them accessible, and make sure your care team has what they need to fight for the imaging you need.
Need to share imaging studies for a prior authorization appeal? Get started with Medixshare — share your scans instantly with any physician, no CD or portal login required.
Looking for faster imaging analysis to support clinical decisions? Explore MYAIRA AI — AI-assisted analysis across 15+ pathologies in under 3 seconds.