📰 Key Takeaways

Insurance companies, through analysis from the Blue Cross Blue Shield Association, found that hospitals using AI tools when submitting insurance claims drove an extra $942 million in healthcare spending over two years. The analysis shows a marked rise in patients coded as having “complex conditions,” without a corresponding change in actual treatment, revealing a “clear disconnect” between medical coding and actual care. The New York Times treats this analysis as the latest evidence that AI is pushing up healthcare costs, noting that while the tug-of-war between hospitals and insurers over treatment and payouts is nothing new, things seem to have gotten worse now that both sides are using AI. Shiv Rao, founder of AI startup Abridge, admits the misuse of AI could lead to “a dystopian future nobody wants to live in,” where it’s “bots fighting bots, agents fighting agents” — though he also believes AI could help ease tensions and cut costs. BCBSA senior VP Luke Chalker, meanwhile, refuses to call it a “war.” He says it’s really “a complete and total bloodbath,” and insurers are on the losing side.

💬 JudyAI Lab Take

Insurance companies, through analysis from the Blue Cross Blue Shield Association, found that hospitals using AI tools when submitting claims drove an extra $942 million in healthcare spending over two years — this story is worth paying attention to because it puts a hard number on the risk of AI being misused for medical coding.

The analysis shows a marked rise in patients coded as having “complex conditions,” with no corresponding change in actual treatment, revealing a clear disconnect between coding and care. This points to a lesson for AI builders: when AI gets used to optimize one side’s metric (like claim amounts) instead of optimizing for real outcomes, the system gets gamed without anyone even noticing. Abridge founder Shiv Rao mentioned that once both sides are using AI, it could turn into “bots fighting bots.” BCBSA VP Luke Chalker describes it as a one-sided “bloodbath” where insurers are losing. This shows tools themselves are neutral, but when both sides weaponize AI against each other’s systems, the problem gets amplified instead of solved.

Something for AI builders to sit with: when designing an AI system, don’t just check whether it hits the target metric — think hard about whether that metric actually equals the “right outcome.”

📅 Original Source Info

🔗 Further Reading