STAT+: Dartmouth put its AI therapy chatbot through the RCT wringer. Is it better than playing Tetris?

You’re reading the web edition of STAT’s AI Prognosis newsletter, our subscriber-exclusive guide to artificial intelligence in health care and medicine. Sign up to get it delivered in your inbox every Wednesday. 

Last week, NEJM AI published a study on a randomized clinical trial showing that a generative AI chatbot improved the symptoms of people with depression, anxiety, and eating disorders on par with gold-standard, human-administered therapy. One psychiatrist online called it a breakthrough so big you might consider it “psychiatry’s ChatGPT moment.” 

But independent experts I talked to shot down that notion. John Torous, director of the digital psychiatry division at Beth Israel Deaconess Medical Center in Boston, said that social media posts are blowing the study — which is indeed an exciting study showing what’s possible — out of proportion. “I saw the post online, I was like, ‘Oh, this is really big,’” said Torous. “And then my face just got, like, sad and angry as I kept reading it.”

Continue to STAT+ to read the full story…