AMA assigns clinical AI safety duties to physicians, developers and health systems

The AMA has issued an issue brief assigning responsibilities for clinical artificial intelligence to physicians, developers and health care organizations, arguing that patient safety requires accountability across the groups. Under the framework, physicians should use clinical judgment, question AI outputs and dismiss recommendations when needed. The AMA says physicians should not be responsible for errors they cannot control.

Developers should provide safe and accurate recommendations and give users information about a tool’s intended use, performance and limitations. Health care organizations should evaluate tools before deployment, train clinicians, monitor performance and establish processes for adverse events. The 2026 AMA survey found that a quarter of physicians received no AI training and 92% were interested in more training. The framework also says patients harmed by AI should have routes to seek recourse, including compensation, according to Goldsack.