AI Is Taking Over Hospitals
SMRTR summary
A chatbot beat hundreds of doctors at diagnosing patients in a landmark Harvard-Stanford study published in the journal Science. But the lead researcher, Adam Rodman, urged caution at the press conference: "I get a little bit queasy about how some of these results might be used."
That caution isn't slowing anyone down. Eighty percent of U.S. doctors are already using AI tools on the job, according to a 2026 American Medical Association survey, and hospitals keep rolling out new AI products, most of which have never been approved by the FDA.
The regulatory gap is real. Many AI health tools skirt FDA oversight by classifying themselves as "clinical decision support," not medical devices. Meanwhile, studies show AI can lead doctors astray with false information, and may fail to flag medical emergencies for patients.
One researcher compares the moment to Uber's rise, when a new industry outpaced regulators entirely. The health-care system, Rodman warns, is not going to "slow down and wait for the evidence to accrue."
SMRTR provides this summary for quick context. The original article belongs to Reddit.
Read the original article