AI-driven denials spark outcry
AI systems in healthcare and claims are being criticized after pilots and deployments reportedly denied or delayed care, with experts warning automated decisioning can disenfranchise patients and require stronger explainability and oversight ( ).
The Centers for Medicare & Medicaid Services launched the WISeR (Wasteful and Inappropriate Service Reduction) Model on January 1, 2026 to test technology‑enabled prior authorization in Traditional Medicare. (cms.gov) The demonstration requires prior authorization or pre‑payment review for select services in six states—Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington—and is scheduled to run through December 31, 2031. (dlapiper.com; federalregister.gov) CMS designated private technology firms (not provider organizations) as the model participants, and public disclosures list participants including Humata Health, founded by interventional radiologist Jeremy Friese, MD. (dlapiper.com; radiologybusiness.com) Local reporting in Washington documented patients facing unmet authorization requests and clinicians reporting new workflow bottlenecks after the program’s January start, with multiple patients described as experiencing untreated pain while awaiting decisions. (seattletimes.com; msn.com) Investigations and oversight bodies have raised incentive concerns: Stateline reported the payment structure could reward cost‑avoidance, and a Senate review found UnitedHealthcare, CVS Health/Signify and Humana denied post‑acute care at substantially higher rates from 2019–2022, noting Humana’s 2022 post‑acute denial rate was roughly 16 times higher than its overall denial rate. (stateline.org; fiercehealthcare.com) Professional groups and lawmakers pushed back: the American Medical Association has cautioned that prior authorization processes continue to disrupt care delivery, and House Democrats introduced legislation aiming to block the WISeR prior‑authorization demonstration. (ama-assn.org; radiologybusiness.com) Meanwhile, InsurTech and industry outlets continue to market AI for claims and intake automation—Coruzant’s March 21, 2026 roundup highlights chatbots handling FNOL, policy questions and fraud detection—and consultancies such as PwC have published use cases for AI across claims intake, fraud detection and predictive modeling. (coruzant.com; pwc.com)
