Hospitals and insurers use AI in billing, and payment disputes grow

In short: Hospitals and health insurers are both using AI in medical billing and claims, and reports say it is adding speed and scale to payment disputes.

What's going on

Hospitals and health insurers are using artificial intelligence, or AI (software that finds patterns in data, like a fast assistant), in the process of billing and paying for care. New reporting says this is intensifying arguments over what care should be covered and how much should be paid.

Hospitals and health systems are using AI tools to draft clinical notes with “ambient scribes” (apps that listen during a visit and write notes), and to pull extra details from patient records. They are also using AI for the revenue cycle, which is the back office work of turning care into bills, including coding, eligibility checks, and sending claims.

Insurers are using AI for prior authorization (asking permission before certain care is covered), reviewing claims, and looking for fraud or unnecessary billing. A Sept. 24 Reuters report cited a Blue Cross Blue Shield Association study that linked provider use of AI to nearly $1 billion in extra costs over two years. Insurers say their own AI tools help catch questionable charges.

The conflict often comes down to documentation versus denial. Hospitals say AI helps them document care more completely and correct “undercoding” (billing at too low a level by mistake), which can raise payment when the care was actually provided. Insurers and their advocates say some provider tools are used to push bills into more complex and higher paid categories.

What to watch

Surveys suggest AI use is already common, including a 2025 survey that found 84% of health insurers used AI or “machine learning” (a type of AI that learns from examples) across products. Regulators, employers, and patients will likely focus on whether AI leads to more incorrect denials, higher costs, or both, since it can automate the old billing fight and make it happen faster.

Source: NYTimes