The Food and Drug Administration has been talking a big game about bringing artificial intelligence to patients. In January, when it announced relaxed rules for certain AI products, Commissioner Marty Makary said the agency is “developing a new regulatory framework for AI.”
How the agency will regulate rapidly-evolving uses of generative AI is one of the big questions facing health technology developers. Large language models’ wide-ranging applications evade simple measures of safety and efficacy, challenging the FDA’s longstanding approach to device validation — and the agency has yet to authorize a device that relies on generative AI. But a recent breakthrough designation from the FDA could offer hints about its approach to regulating patient-facing chatbots that fall under its purview.
In November, the FDA quietly handed one of its breakthrough device designations to a chatbot for patients recovering from joint replacement surgery. Under development by RecovryAI, which is coming out of stealth as it announces the designation, the LLM-powered device would be prescribed to patients to use in the 30 days after surgery. It will encourage them to check in twice a day about their sleep, activity, diet, and other elements of recovery, answering questions and escalating to a care team when necessary.
This article is exclusive to STAT+ subscribers
Unlock this article — plus in-depth analysis, newsletters, premium events, and news alerts.
Already have an account? Log in
