Health care leaders and technology companies are investing heavily in large language models, a powerful class of deep learning models that are trained on vast troves of data. But while enthusiasm for generative AI is high and research is booming, it’s just starting to gain a toehold in clinical settings. Generative AI could be a revolution in health care — but along the way, it will have to address concerns about whether tools meet standards of safety, privacy, equity, and practicality.
To keep an eye on the real-world impact of generative AI on medicine, STAT created a tool to track the uses that are closest to reaching patients, including those that health systems have publicly acknowledged are in pilot or production. The Generative AI Tracker is populated with examples that health systems and vendors have promoted in press releases, along with those surfaced by exclusive STAT reporting and surveys of technology vendors and leading hospitals. The list isn’t complete — many health systems choose not to share their implementations at early, riskier stages — but the tracker will be updated regularly with new additions and product details as they become public.