Artificial intelligence tools are now active on both sides of a familiar struggle in U.S. healthcare - the tension between providers seeking higher reimbursement for medical services and insurers looking for evidence that those services were clinically necessary. The proliferation of AI in documentation, coding and claims review has intensified that conflict, producing contested outcomes and uncertain winners.
Insurer Centene has publicly raised concerns about hospital adoption of revenue-capture software that it says can drive more aggressive coding. "There have been some of these pockets where folks coming into the emergency department with a fever, all of a sudden all have sepsis," Centene CEO Sarah London said at a September investor conference, citing sepsis as an example of a condition that triggers heightened treatment and billing.
A Blue Cross Blue Shield analysis of its commercial hospital claims found substantial sums potentially linked to more aggressive, AI-enabled coding nationwide - roughly $663 million in inpatient spending and at least $1.67 billion in outpatient spending, the insurer reported. "We are seeing more AI tools used at different points in the care and billing process, and when those tools operate independently, they can unintentionally lead to friction," said Razia Hashmi, vice president of clinical affairs at the Blue Cross Blue Shield Association.
How AI is being deployed
Insurers have increased the use of automated systems to screen treatments and claims they judge to be unwarranted. At the same time, hospitals are deploying AI-powered documentation and coding instruments that can result in higher billed amounts by more fully capturing services rendered, according to company statements and interviews with experts and analysts.
Those moves occur against a backdrop in which the United States spends more on healthcare than other nations, at about 18% of gross domestic product. Both providers and payers cite cost control as an objective for their AI investments.
Consultancy McKinsey estimated that for every $10 billion in revenue, AI could save insurers $970 million through improvements in claims management, handling prior authorization requests and by guiding clinical care. Separately, Morgan Stanley said in a research note that AI-driven hospital care savings could reach as much as $900 billion by 2050.
Analysts and industry estimates
Some analysts have projected saving opportunities for both insurers and hospitals using AI. TD Cowen’s Ryan Langston and Whit Mayo of Leerink Partners, among others, have said both sides can realize cost reductions, though they did not publish specific dollar estimates in the disclosures cited.
Spending on healthcare AI has surged, according to Menlo Ventures, a venture capital firm that tracks early-stage investments in AI and health technology. Menlo reported that healthcare AI spending reached $1.4 billion in 2025, nearly triple the level in 2024 based on a survey of 700 industry executives. That tally included about $1 billion from health systems, roughly $280 million from outpatient providers and about $50 million from payers.
Corporate AI plans and savings targets
UnitedHealth Group has cited AI as a potential source of nearly $1 billion in savings in 2026 and said it expects to invest about $1.5 billion in AI this year with at least the same planned for 2027. UnitedHealthcare, the company’s insurance unit, has directed much of its AI effort toward improving the consumer experience, including steering patients to higher-quality care, executives have said.
Smaller rival Humana recently estimated that its AI investments would generate more than $100 million in savings over a span of years. The company declined to provide further detail on those figures. CVS Health said its Aetna insurance business is investing in AI to help improve clinical care and is partnering with providers to ensure appropriate care. Cigna and Elevance did not respond to requests for comment on their approaches to AI in payer-provider interactions.
On the provider side, HCA Healthcare, the largest publicly traded hospital chain in the United States, said in January it expects roughly $400 million in cost savings in 2026 from AI initiatives. HCA has used AI to automate revenue management and to assist with clinicians’ documentation. HCA Chief Financial Officer Michael Marks previously characterized the company’s use of AI as a response "to the growing denial and underpayment activities from the payers." The company did not respond to a request for additional comment.
Providence, a system of 51 hospitals across seven states, said AI tools are helping to more accurately represent the services delivered to patients and thereby enable more exact reimbursement from payers. Providence Chief Health Information Officer Maulin Shah said the introduction of AI will require changes in the dynamic between payers and providers. "It’s going to require adjustments in the relationship between the payers and the providers to understand this new reality," Shah said. "Unfortunately, what we’re seeing is AI fighting AI."
Experts on automation at scale
Christina Silcox, research director of digital health at the Duke-Margolis Institute for Health Policy, cautioned that a conflict of algorithm against algorithm could leave both sides disadvantaged. "The idea of (AI) bot versus bot is intrinsically a situation where no one’s going to win," she said.
Industry observers note that AI is also being widely applied to administrative tasks across healthcare operations, not only to clinical documentation or claims review, suggesting broader implications for efficiency and labor use.
Summary
- AI is being used by hospitals to more fully document services and by insurers to detect and challenge questionable claims, increasing the operational friction between the two groups.
- Analyses and company disclosures show significant sums of spending and potential savings tied to AI, but experts warn that competing automated systems could lead to stalemate rather than decisive advantage.
- Major organizations on both sides - including payers such as UnitedHealth and insurers like Humana, and hospital systems such as HCA and Providence - have reported substantial AI investments and savings targets.
Key points
- Insurer analyses have linked hundreds of millions to billions of dollars in hospital spending to more aggressive, AI-enabled coding practices - roughly $663 million for inpatient and at least $1.67 billion for outpatient claims according to Blue Cross Blue Shield figures.
- Market estimates and corporate plans show rapid growth in healthcare AI spending, with Menlo Ventures reporting $1.4 billion spent in 2025 and UnitedHealth planning multi-hundred-million to billion-dollar investments and savings initiatives.
- Both sides emphasize cost control and operational gains from AI, but the interaction of independent tools across the clinical-to-billing chain has produced unintended friction and requires new coordination between payers and providers.
Risks and uncertainties
- Competing AI tools used independently by providers and payers can generate conflicts that increase administrative friction and uncertainty in reimbursement outcomes - affecting hospitals, insurers and patients.
- Large projected savings from AI are estimates and may not materialize uniformly across organizations or over anticipated timeframes; corporate targets and vendor studies cited different timelines and magnitudes.
- The rapid escalation in AI spending by health systems and uneven investment by payers creates an imbalance in capability and influence that could alter negotiation dynamics but does not guarantee better clinical or financial outcomes.
The deployment of AI in the billing and claims arena marks an intensification of a long-running dispute over what care should be paid for and at what level. As both hospitals and insurers scale their use of automated tools, the industry will have to confront not only the potential efficiency gains but also the operational and contractual frictions that arise when independent systems intersect.