- Published on
Insurers claim AI is already increasing healthcare costs
- Authors
- Name
- DeepHorus
Artificial intelligence at hospitals is adding millions of dollars to U.S. healthcare spending, according to an analysis from the Blue Cross Blue Shield Association. The study estimated that AI-assisted insurance claims contributed to $942 million in additional spending across a two-year period. The finding matters because it turns automationās promised efficiencies into a new front in the long-running struggle over medical coding, reimbursement, and who controls healthcare costs.
The dispute is not new. Hospitals and insurers have long disagreed over how treatments are documented, coded, and paid, particularly when complex cases require nuanced judgments. Electronic health records, billing software, and earlier forms of automation gradually reshaped that process. Generative AI now accelerates the trend by generating clinical notes, identifying billable conditions, and recommending codes at a scale hospitals can no longer easily ignore.
BCBSA researchers reported a sharp rise in patients being documented as having complex conditions during the two years they examined. The association said there was no evidence that the increase corresponded to a change in treatment, pointing to what it called a clear disconnect between coding and care. Its interpretation is that AI systems are capturing more complexity in claims even when the underlying clinical course of care has not changed.
Typical hospital AI tools listen to physician-patient encounters, draft notes, suggest diagnoses, and map those records to billing codes. They may flag possible comorbidities or omissions for coding staff, allowing more services and higher-severity diagnoses to enter a claim. Insurers can also apply machine learning to review claims, audit patterns, and determine coverage, creating what Abridge founder Dr. Shiv Rao described as ābots fighting bots, agents fighting agents.ā
Rao, whose startup develops AI for clinical documentation, warned that the technology could lead to a āhorrible dystopic future nobody wants to live inā if hospitals and insurers deploy competing systems without guardrails. He nevertheless said better-aligned AI could reduce friction, standardize documentation, and lower costs. BCBSA senior vice president Luke Chalker offered a more confrontational assessment: āItās not a war. Itās a completely one-sided bloodbath,ā with insurers bearing the burden.
For providers, richer documentation may improve revenue-cycle performance but can also expose patients to higher bills, premiums, denials, or audits. For insurers, the immediate effect is increased claim payments, while policymakers and employers face pressure to determine whether AI is revealing previously uncoded care or simply inflating reimbursement. The New York Times characterized the analysis as evidence that AI is worsening an existing payment conflict rather than resolving it.
The broader market is likely to respond with tighter coding controls, audit programs, and demands for evidence that documentation changes reflect actual treatment. Hospitals may face pressure to disclose how they use AI, while vendors may need independent testing to measure its effects on coding accuracy, patient safety, and costs. Regulators and professional medical organizations will also need clearer standards before automated systems can be trusted across clinical and billing workflows.
In the near term, the $942 million estimate is likely to intensify negotiations between health systems, insurers, and government payers. Over the longer term, shared standards and interoperable systems could make AI reduce administrative work instead of amplify disputes, but only if incentives are aligned. The next phase of the debate should focus on audited data, transparent methodology, and longitudinal studies comparing AI-assisted claims with actual patient outcomes. Until then, AIās role in healthcare remains a major cost risk.
