Hospitals' AI Coding Tools Helped Add $942 Million in Questionable Charges, Insurer Group Says

AI is moving deeper into the administrative core of the US healthcare system - not just helping doctors diagnose and treat patients, but shaping what hospitals bill, what insurers pay and which claims get denied. Now one of the country's largest insurer groups says that shift is showing up in the numbers. The Blue Cross Blue Shield Association (BCBSA) estimates that hospitals' use of AI-assisted medical coding contributed to close to $1 billion - $942 million - in additional costs for its health plans between 2023 and 2025.

Much of the increase came from secondary diagnoses that moved patients into higher-paying reimbursement categories, the group's analysis found - the kind of diagnoses that "may be derived from single laboratory values, making it particularly well suited for detection by AI tools." The growth in what BCBSA calls "complex coding" occurred during a period when 60% of hospital systems began using AI coding tools. "There is a clear disconnect between coding and treatment," the report stated.

Luke Chalker, BCBSA's senior vice president of product and data science, told CNBC that roughly 70% of the billing identified - about $653 million - was tied to additional diagnoses that were not accompanied by a change in care. He stopped short of attributing the entire increase to AI. "While multiple factors contribute to coding intensity, the findings suggest AI-enabled coding and documentation tools are playing a role," he said, adding that more complex coding without more care "can eventually show up in the form of higher premiums and out-of-pocket costs."

Health economists see the tools amplifying incentives that already existed. "AI appears to be accelerating, and in some sense making it easier to capture, the existing and underlying billing incentives that are in the system," said Christopher Whaley, a health economist at Brown University who studies hospital coding. In many cases the extra diagnoses are legitimate and simply were not captured before, he noted - but others are conditions that "clinically just don't really matter and don't influence the patient's care," while still adding a billable code.

The American Hospital Association pushed back on the analysis. "Patients today are older and more clinically complex," a spokesperson said, adding that AI tools help providers "appropriately capture their patients' conditions to aid in care planning," and that the BCBSA report "lacks the context needed to meaningfully assess how these tools impact healthcare quality, patient access, or spending." The group also accused insurers of relying on their own automated downcoding and denial practices that impede coverage of medically necessary care.

With hospitals, insurers and claims reviewers all deploying increasingly sophisticated AI, economists warn of an "administrative arms race" in a system that is already the most expensive in the world. Benefits consulting firm Marsh forecasts the cost of employer health coverage per employee will rise 8.2% on average in 2027 - the highest increase since 2003. "If providers use AI to improve coding and insurers respond with their own AI tools that lead to more denials, the arms race is poised to exacerbate," said Marisa Greenwald, a partner at Oliver Wyman's health practice. Otherwise, she warned, healthcare could end up with "robots talking to robots and just fighting with each other."

Even industry voices sympathetic to providers draw a line at fully autonomous coding. "The AI is new, but the rest of it is not new," said Vanessa Moldovan, author of "The Healthcare Revenue Cycle AI Playbook," noting that ten human coders can code the same chart ten different ways. Her rule: there should always be a human in the loop, and AI-generated coding should be audited the way human coders' work has always been. Chalker agreed: AI should "support decision-making, not replace human judgment."