> There was actually a project at an innovation center at a well-known medical center which leveraged ML to maximize the amount of codes they could bill for without being rejected. T

I think this perspective makes sense from someone who works on the insurance side of things.

On the other side, there is no way for the insurance company to acknowledge the clinical severity of a patient except via abstruse ICD code choices that only billing clerks know. So this is a perfect case for an LLM - map normal human words onto ICD claim codes to accurately convey patient severity.