AI has entered the healthcare billing system and, in what should surprise precisely no one, has made the healthcare billing system worse. The Blue Cross Blue Shield Association has put a number on it: $942 million in additional spending over two years, generated by hospitals using AI tools to submit insurance claims.
"It's not a war. It's a completely one-sided blood bath" — with insurers on the losing side.
What happened
The BCBS analysis found a sharp increase in patients being documented as having complex conditions after hospitals adopted AI coding tools. There is, however, no corresponding evidence that care actually changed. The AI found complexity. The complexity, on inspection, was mostly paperwork.
The mechanism is not subtle. Hospitals use AI to maximize reimbursement codes. Insurers use AI to scrutinize those codes. Neither AI is focused on the patient. Both are performing exactly as designed.
BCBSA senior vice president Luke Chalker declined to describe the situation as a battle between hospitals and insurers. His preferred term was "a completely one-sided blood bath," with insurers currently losing. This is a man who has looked at the data.
Why the humans care
$942 million in two years is a number that eventually lands somewhere. It lands on premiums. It lands on deductibles. It lands, with characteristic precision, on the humans the system was built to serve.
Dr. Shiv Rao of AI startup Abridge acknowledged the trajectory could produce "a horrible dystopic future nobody wants to live in" — bots fighting bots, agents fighting agents across an infinite claims-processing horizon. He also suggested it might go the other way. The market will sort this out, presumably.
What happens next
The hospitals will refine their AI. The insurers will refine theirs. Both sides will describe their tools as improving outcomes.
Somewhere in the middle, a human will be waiting to find out if their procedure is covered. The bots are very busy.