Artificial intelligence is rapidly becoming the newest weapon in one of U.S. healthcare’s oldest conflicts: how much hospitals charge and how much insurers will actually pay. Hospitals are deploying AI to optimize medical coding and documentation so they can capture higher reimbursements, while insurers are using AI to scrutinize claims, flag questionable services, and deny or reduce payments unless they see evidence that care was necessary. Experts interviewed said the escalation is making it harder to predict who “wins,” and it could raise tensions—and possibly costs—before any savings materialize.
On the insurer side, the concern is that AI-enabled coding tools can push bills upward by identifying diagnoses or billing codes that trigger higher reimbursement. Centene CEO Sarah London pointed to patterns that raised red flags—for example, cases in which patients arriving at emergency departments with routine symptoms were suddenly coded as having severe conditions like sepsis, a diagnosis that typically prompts more testing and treatment and can yield higher payments. Insurers view these changes as “aggressive” or potentially improper use of revenue software.
Blue Cross Blue Shield offered a concrete estimate of how big the impact could be. Its analysis of commercial hospital claims suggested that roughly $663 million in inpatient spending and at least $1.67 billion in outpatient spending could be linked to more aggressive, AI-enabled coding practices nationwide. The scale of those numbers helps explain why insurers are investing heavily in countermeasures, including AI systems that review claims and validate the clinical rationale for services.
Insurers argue that controlling these costs is becoming essential as medical expenses climb and profit margins get squeezed. UnitedHealth, Humana, CVS Health’s Aetna, and others are investing in AI—often promising big savings from automating processes and steering members to appropriate, high-quality care. Humana, for example, has estimated AI investments could generate more than $100 million in savings over a few years.
Hospitals, meanwhile, say they’re using AI partly in self-defense. Providers argue that denials and “underpayment” have increased, and they need better tools to ensure the care they deliver is fully documented and reimbursed. In fact, HCA Healthcare, the largest publicly traded U.S. hospital chain, has said it expects about $400 million in 2026 cost savings from AI initiatives, including automating revenue management and reducing clinicians’ paperwork. Providence, a large nonprofit system that AI can help “accurately represent” services rendered—essentially making sure the billing matches the complexity of patient care.
The result, as one hospital executive put it, is a new reality where “AI is fighting AI.” That dynamic could intensify the administrative arms race: hospitals refine coding; insurers refine denials; both add layers of automation; and the system risks becoming more complex even if it becomes faster. Overall, AI spending surged in 2025 to $1.4 billion with hospitals accounting for about 75% of that amount—evidence of how quickly the technology is spreading through billing and payment workflows.
Some analysts believe AI could eventually reduce costs—McKinsey estimates significant potential savings for insurers, and long-range forecasts envision major efficiency gains for hospitals. But the near term will look more like escalation than peace: faster disputes, more automated pushback, and a higher-stakes battle over what care is worth—and who gets to decide.
