The AI That Decides Who Dies (90% Error Rate)

Описание: It’s 2:30 AM. A contractor wakes up a recovering patient to kick them out of the hospital. The doctor says they need to stay. The algorithm says they have to leave.

Five days later, the patient is dead.

In this investigation, we expose the secret AI systems—like UnitedHealthcare's "nH Predict"—that are systematically denying care to elderly and vulnerable patients. When families fight these decisions in federal court, they win 90% of the time. The algorithm is wrong 9 times out of 10.

But the companies keep using it. Why? Because they know only 0.2% of people have the money to fight back.

We analyze the internal documents revealing why insurance executives refer to your medical care as "loss," how State Farm’s algorithms are accused of modern-day redlining, and the specific mechanism that turns patient denials into record profits.


Sources & Research Cited:

🟢Lokken v. UnitedHealth Group: Federal class action lawsuit regarding the use of nH Predict.

🟢Huskey v. State Farm: Lawsuit alleging racial discrimination in algorithmic claims processing.

🟢STAT News Investigation: "Denied by AI: How Medicare Advantage plans use algorithms to cut off care."

🟢Senate Permanent Subcommittee on Investigations: Inquiry into Medicare Advantage denials.

🟢Journal of the American Medical Directors Association: Research on the overturn rates of algorithmic denials.

About Grey's Capital: We analyze the algorithmic economy, institutional finance, and the systemic risks that standard financial news ignores. The machines are already running the world—we just explain how.