Insurers Say AI Is Driving Up Health Care Costs by Nearly $1 Billion

Artificial intelligence tools used in hospital billing and insurance claims appear to have added nearly $1 billion to U.S. health care spending over two years, according to a new analysis from the Blue Cross Blue Shield Association. The insurer group says AI-assisted coding is helping push patients into more complex billing categories without a matching increase in treatment.

The finding adds fresh fuel to a growing debate over whether AI is improving health care efficiency or simply automating new forms of escalation between hospitals and insurers. The Blue Cross Blue Shield Association says the extra spending reached $942 million across a two-year period, making the issue more than a theoretical risk for the industry.

At the center of the dispute is a familiar part of the American health system: medical billing. Hospitals and clinicians submit claims to insurers using diagnostic and procedure codes, and those codes help determine what gets paid. If AI tools are being used to optimize that process, insurers argue, they may also be making it easier to justify higher payments without corresponding changes in care.

What did the Blue Cross Blue Shield analysis find?

The Blue Cross Blue Shield Association says its review detected a large jump in claims that documented patients as having more complicated conditions, while the actual care provided did not show a parallel rise. In the association’s view, that gap points to a growing mismatch between the billing language on claims and the medical services patients are receiving.

According to the analysis, the pattern translated into an estimated $942 million in additional health care spending over two years. That figure is the core of the group’s warning: AI is not only changing how hospitals work, it may be changing what they get paid for.

The association did not frame the problem as a case of intentional fraud. Instead, it described a sharp change in documentation patterns and argued there is no evidence that the underlying care improved or changed enough to justify the rise in costs.

Why are insurers worried about AI in claims processing?

Insurers are worried because AI can make it easier and faster to generate documentation that supports higher reimbursement. In a system already built on narrow definitions, code selection, and constant negotiation over payment, even small shifts in coding behavior can produce major financial consequences.

That concern is not limited to one side of the market. The New York Times, which highlighted the analysis, described the development as another sign that AI is affecting health care costs rather than lowering them. The publication said long-running fights between hospitals and insurers are now being sharpened by AI tools on both sides.

Hospitals have strong incentives to make sure patient severity is documented accurately, especially when treatment is expensive and reimbursement depends on the level of complexity recorded in the claim. Insurers, meanwhile, are watching for signs that documentation may be drifting beyond what the record can support.

The Blue Cross Blue Shield Association’s view is that AI may be helping widen that gap. If software can rewrite, summarize, or optimize claim language at scale, then claims may look more severe on paper even when the underlying care remains unchanged.

How could AI raise health care costs?

AI could raise health care costs by improving a provider’s ability to code claims at higher levels, detect wording that improves reimbursement, and flag language that insurers are likely to reject. Those gains may not change medical care itself, but they can change how much a system pays for it.

That dynamic matters because the billing process is not a side issue in American health care. It is one of the main mechanisms through which money moves from insurers to providers. When AI enters that workflow, it can influence spending even if it never touches a patient bedside.

Experts and industry participants have long worried that machine-generated paperwork could create an arms race between institutions trying to automate their advantage. This latest analysis suggests that the concern is no longer hypothetical.

Are hospitals and insurers entering a bot-vs-bot era?

They may be, although not everyone thinks that is the right way to describe the problem. Hospital systems are increasingly using generative AI to help draft notes, summarize visits, prepare claims, and reduce administrative work. Insurers are also adopting AI to review submissions, identify inconsistencies, and decide what should be paid.

That shift means more of the claims process is being filtered through software. Instead of a human reviewer on one side and a human biller on the other, the decision chain may increasingly involve models that can produce, compress, or challenge documentation at speed.

Dr. Shiv Rao, who founded the health AI startup Abridge, said the industry could end up in a grim “bots fighting bots, agents fighting agents” environment if the technology is used purely to game payment systems. He also suggested, however, that AI could reduce friction and ultimately lower costs if deployed more constructively.

Rao’s view reflects a wider uncertainty in health care AI. The same tool that can remove paperwork burden can also make it easier to push the system harder. Whether the result is efficiency or escalation may depend on how aggressively each side adopts it.

What did the insurers say about the situation?

The Blue Cross Blue Shield Association’s senior vice president, Luke Chalker, rejected the idea that the sector is locked in an even contest. His message was that insurers are already losing the struggle to keep costs in check.

Chalker argued that this is not a balanced fight, describing the situation as overwhelmingly one-sided in favor of hospitals and providers rather than insurers.

The forcefulness of that language underscores how contentious the issue has become. Insurers are not simply saying that AI may create errors; they are arguing that the technology may be helping providers gain the upper hand in payment negotiations and claim submissions.

That stance also reflects a broader political and economic reality. Any increase in claims spending eventually shows up in premiums, employer health costs, and patient out-of-pocket expenses. Even if a higher bill begins with an AI-assisted coding decision, the ripple effects are felt far beyond the hospital finance office.

Why this matters beyond one analysis

This is not just a story about one insurer group’s review. It is a signal that AI’s influence in health care is moving deeper into the financial machinery of the system, where code selection and documentation can be as consequential as the treatment itself.

If the analysis is correct, then AI is not only automating administration. It is also altering the economics of care delivery in a way that could push costs up rather than down. That makes the technology a policy issue, a business issue, and a consumer issue at the same time.

For patients, the concern is indirect but real. Higher spending in the insurance system typically does not stay contained within provider contracts. It can affect premiums, deductibles, network decisions, and the willingness of insurers to approve certain services.

For hospitals, AI can still be attractive because it may reduce paperwork and improve compliance. But the Blue Cross Blue Shield analysis suggests those efficiencies may come with a new risk: if coding becomes more aggressive, every productivity gain could be offset by a higher reimbursement battle.

How the health care AI debate has changed

For much of the public discussion around AI in medicine, the focus has been on patient-facing tools, clinical decision support, and administrative burden. This report shifts attention to the middle of the revenue cycle, where the financial incentives are often strongest and the oversight is less visible.

That is important because claims processing has always been a point of conflict between hospitals and insurers. AI may not have created the tension, but it may be making the consequences more expensive and the pace of escalation faster.

The New York Times characterized the trend as another indication that AI can contribute to rising medical costs. That broader framing matters: health care is one of the few sectors where automation can plausibly increase output, speed, and complexity at the same time.

Key timeline of the issue

The current debate is unfolding quickly, but it sits on top of years of pressure on billing and reimbursement. The table below summarizes the immediate context behind the latest report.

What happens next?

The immediate next question is whether other insurers, regulators, or researchers will find similar patterns in their own claims data. If they do, the debate will likely shift from speculation to enforcement, oversight, and possible policy changes around the use of AI in billing workflows.

Hospitals, for their part, are likely to push back on any suggestion that AI-driven documentation is inherently suspicious. Many will argue that better coding can simply reflect better record-keeping, more complete notes, and a more accurate picture of patient complexity.

That is the core tension in the story. AI may be exposing inefficiencies in a broken billing system, or it may be inflating them. The answer could differ by institution, by tool, and by how aggressively providers use the software.

What is clear is that AI has moved from the clinical frontier into the financial plumbing of health care. In that environment, even small changes in documentation can carry enormous cost implications.

For now, the Blue Cross Blue Shield Association’s $942 million estimate gives the debate a hard number — and a warning that AI’s most immediate effect in health care may be to make the system more expensive, not less.