A new study by researchers at Duke University looks at the bumpy rollout of AI in healthcare systems, and describes what it’ll take to make AI into a useful tool for healthcare providers.

Meanwhile, a study in JAMA Internal Medicine found that ChatGPT “decisively bested doctors at providing high-quality, empathic answers to medical questions people posed on the subreddit r/AskDocs.” The responses were judged by a panel of three physicians with relevant medical expertise.

(Ars Technica / Beth Mole)