I presume the insurance providers thought they would be the only ones using the AI tools to auto-process (find creative ways to reject) insurance claims.

Even just the fact that people might learn about their rights, or find better ways to express their issues with correct wording would yield higher approvals, which would mean lesser profit margins for providers. And of course they will lobby for additional government subsidies.

A health insurance cartel is asserting that AI-enhanced diagnostic assistance tools (such as medical record scanning or ambient voice transcription during consultations) is resulting in an increase in diagnosed "secondary conditions" without a corresponding increase in treatment of those conditions. Despite being untreated, these secondary conditions still cost the insurers money because they allow doctors to label a medical case as "complex".