AI billing tools are inflating healthcare costs by documenting more details that justify higher-paying billing codes, even when actual patient care hasn't changed.

What happened: A PwC report identifies AI as one of five drivers expected to push U.S. healthcare costs up to 9% in 2027. The mechanism: AI note-taking tools are capturing more granular details about diagnoses and complications, which hospitals then convert into higher-severity billing codes—standardized labels that tell insurers what to pay. A Blue Cross Blue Shield analysis found that one hospital system saw billing codes for acute posthemorrhagic anemia in maternity cases jump from 4% to 12.3% of admissions between 2022 and 2025, yet the actual number of blood transfusions (the standard treatment) barely increased. An audit revealed fewer than 20% of those cases met clinical criteria for the diagnosis. This "coding intensity" added $22 million（約35億円） to maternity spending at the studied hospitals over three years.

Why it matters: Healthcare costs are already a significant financial pressure on patients and insurers. If AI's primary near-term effect is to justify higher billing rather than reduce waste, it undercuts the industry's expectation that AI would lower costs through automation or efficiency. The report shows that companies are using AI not to streamline care, but to optimize billing in their favor—a reminder that technology amplifies incentives rather than neutralizing them.

What to watch: While AI is the report's top-ranked new cost pressure, labor and supply costs still account for more of the overall increase. The report's authors note that AI tools could eventually drive costs down by automating administrative work or catching diagnoses earlier—so the trajectory is not yet locked in, but current hospital deployment patterns suggest cost-boosting applications are arriving first.