Fitch: Regulators Will Keep Humans in AI Denial Decisions
In its Oct. 5 AI Stress Test report, Fitch Ratings assesses how adverse AI scenarios could affect credit across 107 sub-sectors, scoring health insurers at 20 out of a 40-plus threshold that would signal potential negative rating action, indicating limited near-term disruption risk. Fitch finds that while AI is speeding up prior authorization and other administrative functions, regulatory and legislative scrutiny has already prompted reductions in prior authorization requirements for some services, and expectations for human review of care-denial decisions are likely to persist, partially offsetting efficiency gains. The report also describes an escalating AI arms race between providers optimizing coding for reimbursement and insurers deploying AI to flag aggressive coding and price for it. Fitch has not taken any rating actions tied to AI disruption but says large, diversified insurers face less competitive risk than those concentrated in specific markets.