Insurers Say Hospitals’ AI Use May Already Be Raising Healthcare Costs

Artificial intelligence may be changing more than clinical workflows in the US healthcare system. It may also be changing how hospitals describe patients and submit claims to insurers. An analysis by the Blue Cross Blue Shield Association (BCBSA) says hospitals’ use of AI tools during the claims process was associated with an additional $942 million in healthcare spending over a two-year period.

Key points

• Complex-condition coding rose sharply. BCBSA said the number of patients documented as having complex conditions increased substantially during the period examined.
• The coding did not appear to track care. The association described a clear disconnect between medical coding and treatment, saying it found no evidence of a corresponding change in care delivered.
• AI may intensify the payer-provider struggle. Hospitals can use AI to draft, complete, or optimize claim documentation, while insurers can deploy automated systems to review, challenge, or deny claims. A longstanding payment dispute could therefore become faster and harder to inspect.
• The evidence needs careful qualification. The material available describes an insurer-backed analysis. It does not establish that AI directly caused every dollar of the increase, nor does it provide the full methodology, sample details, or causal analysis needed to settle that question.

Why it matters

Medical coding is not merely clerical. It can affect risk adjustment, reimbursement levels, coverage decisions, and the amount ultimately paid. If an AI system identifies additional diagnoses, fills gaps in documentation, or recommends more complex codes, the record may become more severe even when the underlying treatment has not changed. Hospitals may view this as better documentation, while insurers may see it as an unjustified increase in risk and cost.

The larger concern is that AI could move payment disputes from human review into automated opposition. Shiv Rao, founder of the AI startup Abridge, acknowledged the possibility of a dystopian future of “bots fighting bots” and “agents fighting agents.” He also argued that well-designed systems could reduce tension and lower costs. Luke Chalker, a senior vice president at BCBSA, described the situation more forcefully, saying it was not a war but a one-sided bloodbath in which insurers were losing.

The episode should not be reduced to the claim that AI inherently makes healthcare more expensive. The central issue is whether the industry can distinguish clinical value from documentation efficiency and payment optimization. That requires auditable coding histories, meaningful human review, and independent checks on unusual increases in complex-condition coding. If AI merely makes it easier for two billing systems to compete, spending could rise. If it improves transparency, verifies medical necessity, and removes repetitive administrative work, the outcome could be different. For now, the BCBSA analysis is best read as a governance warning rather than a final verdict on AI in healthcare.

Source: TechCrunch AI