The news: Hospitals' use of AI to document billing codes drove an estimated $942 million in additional healthcare spending for Blue Cross Blue Shield plans between 2023 and 2025, per a Blue Cross Blue Shield Association (BCBSA) analysis. The study used de-identified claims data from BCBS companies, whose coverage of one in three Americans spans every US ZIP code.

Digging into the details: Spending rose because AI catches secondary health issues alongside a patient's main diagnosis. Secondary diagnoses accounted for about 70% of the additional costs, or $653 million, with hospitals receiving an average of $11,800 for each complex case. AI can flag conditions based on structured, objective data, rather than a person’s judgement. For example, if routine bloodwork shows low hemoglobin and other abnormal red blood cell measures, AI can flag possible anemia based on those numbers rather than relying on a clinician to notice symptoms.

Why it matters: Billing for secondary diagnoses is standard practice when patients have those conditions and receive care for them. But the report found AI was surfacing more billable conditions without evidence that patients were sicker. BCBSA noted that anemia diagnoses following major bowel surgery rose sharply at hospitals during the studied period, without a matching rise in blood transfusions, the standard treatment for low red blood cell counts. The report attributed the gap to AI detecting more billable conditions, rather than greater patient illness.

Hospitals, for their part, say AI is improving documentation and cutting time spent on manual charting. By automating transcription and matching health issues to billing codes, the tools help providers document care more accurately and get paid for what they deliver. Documentation of billing codes, medical charts, and visit notes ranks as a top AI use case for physicians, trailing only research summaries and creation of patient care plans, per a March 2026 American Medical Association survey.

Implications for the healthcare system: This is the second BCBSA analysis tying AI coding tools to higher reimbursements without a subsequent rise in treatment. That consistency makes the finding harder for hospitals to dismiss as an isolated coding quirk.

The dispute sits inside a broader standoff between payers and providers over AI. Hospitals accuse insurers of using AI to speed up denials, including for medically necessary care. Neither side has reason to slow its AI investment, since each expects the technology to improve its own margins. But what benefits individual organizations financially may not lower costs across the healthcare system. Many researchers and industry experts expect AI to increase overall healthcare spending even as organizations use it to cut costs and improve efficiency. Those savings may ultimately boost individual organizations’ bottom lines without reducing costs for the system—or patients.