Hospitals use AI to find more things to bill for. Insurers use AI to deny them.

The Blue Cross Blue Shield Association reports that insurers paid almost $1 billion more than typical in 2024-2025 because hospitals used AI to augment documentation. Machine tools scanned records and surfaced secondary diagnoses that human reviewers had missed, increasing coded comorbidities and raising reimbursements despite no rise in treatment intensity. BCBSA’s SVP of data science summarized the effect bluntly: AI is “identifying more billable conditions, not sicker patients.” Insurers have countered by deploying their own AI systems to scrutinize and deny those same claims.

The practical result is an administrative arms race: hospitals buy AI to boost billable codes, insurers buy AI to push back, and AI vendors on both sides capture fees while patients shoulder higher premiums. A large sum shifted between payers and providers without improving health outcomes; the notable change was larger invoices, more denials, and increased revenue for technology vendors and billing departments. This illustrates how automation can amplify existing financial incentives in healthcare, turning technical capability into another layer of cost rather than a tool that directly benefits patient care.