AI billing tools are costing insurers nearly $1 billion — without changing patient care

Hospitals are increasingly using AI to code patient claims as more medically complex — and it's costing insurers big. A new Blue Cross Blue Shield Association (BCBSA) analysis found that the share of medically complex inpatient cases billed to its member plans rose from 37% in early 2023 to 40% by end of 2025, adding an estimated $942 million in costs over two years. Critically, BCBSA found no evidence that patients were actually sicker or receiving more intensive care.

The trend is being driven by widespread adoption of AI revenue cycle management tools, which help hospitals document more diagnoses and code for higher-severity cases — boosting reimbursements even when clinical care stays the same. A separate PwC report found that nearly 70% of health plans ranked AI coding tools as a top-three cost inflator, contributing to a projected 9% rise in commercial healthcare costs in 2027 — the highest in nearly two decades.

By the Numbers:

• $942M in additional costs to BCBSA plans over two years
• $653M of that stemmed from secondary diagnoses (~$11,000 per excess complex case)
• 55,000+ additional cases pushed into higher-severity billing categories
• 9% projected rise in commercial healthcare costs in 2027
• ~70% of health plans cite AI coding tools as a top-three cost inflator

Why it matters: This AI-driven shift in coding practices is quietly driving up premiums and out-of-pocket costs for patients, employers, and taxpayers — without delivering better care. As payers and policymakers scrutinize AI's role in healthcare spending, hospitals may face increasing pressure to demonstrate that higher-complexity billing reflects genuine patient acuity.