Artificial intelligence is supposed to make healthcare faster and more efficient, but new data suggest it may also be driving up medical bills. Hospitals are increasingly using AI-powered tools to help identify and code patients’ medical conditions, and that shift may have contributed to nearly $1 billion in additional insurance costs. The technology could help hospitals capture diagnoses that might otherwise go undocumented, but it also raises an important question: Are patients and insurers paying more because of AI? Read on to find out how hospital AI is changing healthcare costs.
AI is changing how hospitals document patients
Artificial intelligence is becoming a bigger part of hospital operations, from reviewing medical records to transcribing conversations between doctors and patients. But a new analysis from the Blue Cross Blue Shield Association (BCBSA) suggests these tools may be having an unexpected effect: driving healthcare costs higher.
According to the BCBSA, insurers paid nearly $1 billion more in 2024 and 2025 than they did the previous year in connection with AI-assisted hospital documentation and other AI-related work.
The BCBSA represents 31 independent Blue Cross Blue Shield insurers and serves more than 100 million Americans. Its findings suggest the added costs aren’t necessarily coming from the computing power needed to run AI systems. Instead, the technology’s ability to quickly process enormous amounts of medical information could be changing how patients’ conditions are documented and billed.
AI could be finding more billable conditions
A patient may arrive at a hospital with one primary health problem but also have other conditions or complications documented in their medical records. AI tools can scan large amounts of information, including records and patient conversations, and may identify these additional diagnoses more consistently than a human reviewer.
From a medical perspective, finding previously overlooked conditions can potentially be useful. But the U.S. healthcare payment system also takes diagnoses and the complexity of a patient’s case into account when determining reimbursement.
That means documenting more conditions can result in larger insurance payments, even when those additional diagnoses don’t necessarily lead to more treatment.
The BCBSA says this creates a notable gap between what is being documented and what is actually happening during patient care.
“The disconnect between diagnoses and treatment suggests that AI is identifying more billable conditions, not sicker patients,” SVP of Product and Data Science Luke Chalker said.
More diagnoses don’t appear to mean more treatment
The distinction is important because a rise in diagnoses would normally be expected to come with a corresponding increase in treatment if patients were genuinely becoming sicker.
Instead, the BCBSA’s analysis points to a different possibility: AI may simply be making it easier for hospitals to identify and document conditions that can affect reimbursement.
That doesn’t necessarily mean the diagnoses are incorrect. Rather, the technology may be surfacing conditions that were already present but previously received less attention during the documentation and coding process.
The result could be a healthcare system where AI changes not only how quickly information is processed, but also how much insurers ultimately pay for hospital care.
Insurers are using AI, too
Hospitals aren’t the only players bringing AI into the healthcare system. Insurance companies are increasingly using their own AI tools to examine claims and evaluate whether treatments meet medical-necessity requirements.
That creates an unusual dynamic in which AI systems on both sides of the healthcare system are analyzing the same patients and claims for different purposes.
Hospitals can use AI to find and document additional diagnoses, while insurers can use AI to scrutinize claims and determine whether the associated treatments should be covered.
The growing use of AI across both sides could increase computing demands and add another layer of complexity to an already expensive healthcare system. The bigger question is whether all of this additional AI-driven analysis is ultimately producing better patient care — or simply increasing the amount of money moving through the system.
Source:
Techradar