Gen AI policy is a White House priority, CDRH digital health leader says - MedTech Dive

Rick Abramson, director of the FDA's Digital Health Center of Excellence, has taken to the stage at AdvaMed's The MedTech Conference — on a Tuesday, no less — to announce that draft guidance on generative AI will arrive in "the coming months." The coming months. Not the coming weeks, obviously; nothing regulatory happens in weeks. The coming months is the tense in which all American healthcare AI policy lives, a permanent grammatical waiting room with magazines older than the seats. The MedTech Conference got a timestamp; the guidance did not.

The comment period on the FDA's discussion paper on generative AI closed on Monday, which is a lovely bit of choreography: the public gets its say, the window shuts, and the next day a gentleman at a conference explains that the whole thing is a White House priority. He didn't say what the comments said, mind. The discussion paper itself leans on "postmarket monitoring," which is the regulatory term for watching what happens to patients after a model has already been waved through. Very much the smoke alarm installed after the fire, and billed as innovation in safety.

One does admire the ambition of governing a technology that changes quarterly with a document that will arrive in "the coming months," following a comment period that ended Monday, under the umbrella of a White House priority announced at a medtech trade conference. The FDA's device centre is named in the headline as the source of the reassurance, which is the correct place for it: gen AI in a diagnostic is a device until somebody decides it's a service, and that decision is also due in the coming months. Meanwhile every hospital procurement team in the country is already being sold the thing. Postmarket monitoring will have plenty to monitor.