Medicine's first autonomous AI offers a case study in real-world use of tech


nce feared as a job-erasing technology devoid of clinical nuance, autonomous artificial intelligence in medicine is growing up. The Food and Drug Administration in 2018 gave its stamp of approval to the first automated AI screening system, an algorithm that can analyze retinal images to detect diabetic retinopathy. And starting Jan. 1, primary care doctors across the country can get paid more reliably for automated screenings of the vision-threatening condition, which ultimately impacts more than half of people with diabetes.

“That, in my opinion, is going to be a sort of game-changing moment,” said Aaron Lee, an associate professor of ophthalmology at the University of Washington, as a national Medicare reimbursement rate helps more primary care practices decide whether the technology is worth investing in.