Hospitals' use of AI while preparing insurance claims contributed an additional $942 million in healthcare spending over two years, according to an analysis by the Blue Cross Blue Shield Association. The insurer group says patient records showed a sharp rise in complex-condition coding without evidence of a corresponding change in treatment.
Coding determines how diagnoses and procedures translate into payments. Tools that capture more detail can correct omissions, but they can also raise the billed severity of a case. The association argues that the observed gap between coding and delivered care indicates higher payments rather than better treatment.
Insurers are deploying their own automated systems, including tools that review or reject claims. That creates the prospect of providers' software optimizing documentation while payers' software challenges it. Abridge founder Shiv Rao said the result could become “bots fighting bots,” although he also argued that well-designed systems might eventually reduce friction and cost.
The $942 million figure comes from an insurer association with a direct financial stake in reimbursement disputes. The analysis identifies an association between AI-supported coding and spending, but the report does not establish that every additional code was inappropriate. Independent validation and patient-level outcomes would be needed to separate recovered underbilling from unjustified charges.