Many Ideas to Approve FDA Regulation of AI: Article
Politico says that while concern about FDA regulation of artificial intelligence (AI) tools in medicine is growing, there is no shortage of people with ideas for improving the regulation. An online column cites a study by University of North Carolina researchers who looked at 521 FDA-authorized AI medical devices and found that 42% lacked published clinical validation data.
“Patients and providers need a gold standard indicator of efficacy and safety for medical AI devices,” Politico says the researchers wrote. The researchers said FDA should publish more clinical validation data and should prioritize prospective studies.
The article reports potential solutions from people such as former FDA commissioner Scott Gottlieb; Dr. Mark Sendak, an HHS consultant to the national coordinator for health information technology; former FDA director of the Digital Health Center of Excellence Bakul Patel, who is now at Google; and Harvard University public health professor David Blumenthal.
FDA critics say the agency is not prepared to regulate generative AI, which asks and responds to questions, the post says. Some of those who spoke with the news service said generative AI should be regulated less like medical devices and more like clinicians. Patel and Blumenthal said the tools should take required training and tests to evaluate their capabilities, operate under supervision, and take continuing education courses when new scientific evidence emerges. They should be licensed and should report to regulators on the quality of care they are delivering, they said.