A new JAMA Viewpoint by Michelle Mello, JD, PhD, a professor of health policy and of law, examines what AI regulation and governance may look like in the coming years.

“In the US, health and health care artificial intelligence (AI) tools grow in a liminal regulatory space. Wedged uncomfortably between the ambits of regulatory agencies, they are often unreachable by laws governing medical products, information privacy, research, and quality of care. The sparseness of formal regulation contrasts starkly with AI’s potential to cause harm. How and when will this tension be resolved?”

The Outlook for Federal Regulation

“In its first few weeks, the Trump administration has reduced rather than expanded federal oversight of AI,” write Mello and coauthor, Glenn Cohen, of Harvard Law School. “President Trump rescinded a 2023 executive order calling for a range of federal and private activities to strengthen AI governance,” and the mass firing of staff in the FDA’s digital health unit “indicates that stronger AI oversight is not commensurate with the administration’s goal of shrinking federal agencies.”

Filling the Gap

Instead of more federal regulation, the authors suggest that in the near future, we will see greater use of litigation and private contractual agreements between AI developers and healthcare organizations to determine who is responsible for healthcare AI errors. They also comment on the rising role of state regulation—including some state laws that aren’t likely to help avert AI-related harms.

Read the Full Viewpoint