There’s another reason to scoff at AI: It’s hiking up your health insurance costs.

AI has already found plenty of ways to make life more expensive. Data centers are putting pressure on electricity bills, while the hardware feeding the AI boom has helped squeeze cheaper gadgets. Now healthcare may be joining that list.

A new analysis from the Blue Cross Blue Shield Association (BCBSA) estimates that rising hospital coding intensity added about $942 million in healthcare spending for its companies between 2023 and 2025. Medical coding, in simple terms, turns diagnoses and treatment details into standardized codes that are used to calculate hospital reimbursements.

BCBSA says the increase comes as more hospitals use AI-powered coding tools that can scan medical records, doctors’ notes, and lab results for additional conditions that can be added to a claim.

AI can turn the same hospital stay into a pricier claim

Say an AI tool spots anemia, low sodium levels, or another secondary condition in a patient’s records. Adding that diagnosis could move the hospital stay into a more complex billing category, which means a bigger reimbursement for the hospital. BCBSA looked at more than 55,000 additional cases classified as complex and says they generated $653 million in extra reimbursements, or about $11,800 per case.

The concern is that the care itself did not always change. BCBSA found diagnoses being added to claims without a corresponding increase in the treatment normally associated with those conditions. Patients are not simply handed an extra $11,800 bill. The figure refers to additional reimbursement paid to the hospital. Depending on the insurance plan, a pricier claim could mean a larger coinsurance payment for some patients, while others may see no immediate difference. BCBSA says higher healthcare spending can ultimately put pressure on premiums and other out-of-pocket costs.

Hospitals say there’s more to it

The American Hospital Association disagrees with the suggestion that rising coding intensity necessarily means hospitals are inflating claims. It says hospitals are caring for older and sicker patients, while AI tools can also improve documentation and coding accuracy.

BCBSA has evidence supporting its concerns. Among hospitals with the fastest coding growth, more bowel procedures were classified as complex even though ICU use, transfusions, reoperations, and hospital stays were similar or lower.

For now, there is still room for debate over how much of the increase comes from better documentation and how much comes from more aggressive coding. What is clear from BCBSA’s analysis is that insurers are paying more, and some of those higher costs could eventually find their way into what patients pay for healthcare.