Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Insurers Ai topic
No spam. Unsubscribe anytime.
Insurers describe AI for claims processing and fraud detection, stress that humans make adverse determinations
Summary
Insurer witnesses said AI is used for claims OCR, fraud detection and call‑center automation and can speed prior authorization, but they emphasized that adverse benefit determinations are made by clinicians and governance — not automated denial decisions.
Representatives of Highmark and the Insurance Federation of Pennsylvania described payer use cases for AI, from optical character recognition for faxed claims to experimental image‑authenticity tools that aim to flag AI‑generated medical images.
Michael Barber of Highmark said the company uses AI for fraud and waste identification, claims intake automation and member service chatbots. He described an ambient‑listening pilot that can cue prior‑authorization submissions for clinicians to approve, and he stressed that the insurer does not auto‑deny claims: "We don't auto deny anything at all," Barber said; adverse decisions, he noted, are reviewed by a medical director under existing prior‑authorization framework.
Jonathan Greer and Megan Barber of the Insurance Federation framed AI as a tool to reduce administrative burdens and prior‑authorization times, and they cited NAIC engagement on responsible AI practices. Megan Barber said that, with responsible implementation, prior‑authorization processing times have fallen in some use cases from about a week to less than a day.
Legislators pressed for measurable savings and for clarity about vendor liability. Insurers said they include strict contractual limits on vendor use of data and require indemnification clauses, but acknowledged that very large vendors may resist certain contract redlines and that layered vendor relationships complicate accountability. Committee members signaled that liability and vendor accountability will be a subject for future hearings.
