FDA wants feedback on how to regulate generative AI in medicine

August 19, 2026 - 00:07

The U.S. Food and Drug Administration is opening a formal comment period to gather perspectives on how it should regulate generative artificial intelligence used in medical devices. The move signals a major step toward establishing a framework for technologies that can create new content, such as imaging reports, treatment suggestions, or patient summaries, rather than simply analyzing existing data.

Radiologist Rick Abramson, MD, framed the challenge in a statement, saying that generative AI-enabled medical devices are set to reshape the health technology landscape. He added that the FDA has an important responsibility to provide thoughtful leadership for this new era. The agency is specifically asking for input on how to evaluate the safety and effectiveness of these tools, which differ from traditional algorithms because they can produce novel outputs that may change over time.

Key questions include how to handle continuous learning, where an AI model updates itself after deployment, and how to ensure transparency when a clinician relies on AI-generated text or images. The FDA also wants to hear about potential risks, such as bias in training data or the difficulty of validating outputs that are not always reproducible.

Industry groups, healthcare providers, and patient advocates are expected to weigh in over the next several months. The feedback will likely shape future guidance documents and could influence how quickly these products reach clinical settings. For now, the agency is signaling that it wants to move carefully, balancing innovation with patient safety. The comment period is open to the public, and the FDA has not set a firm deadline for final rules.