The inherent risks of AI necessitate ongoing oversight and governance, and physician expertise can’t be substituted, AMA says.
By Jeff Lagasse , Editor – November 17, 2025 –
The American Medical Association is taking a cautious approach to artificial intelligence implementation in healthcare, suggesting to lawmakers that it would be prudent to prioritize education, physician oversight and data security.
The AMA refers to AI as “augmented intelligence” to emphasize that the technology should enhance human intelligence, not replace it.
One of the AMA’s central policy recommendations is that physicians remain at the forefront of decision-making when it comes to AI adoption and be actively engaged in reviewing and validating AI outputs to maintain both clinical accuracy and patient safety.
According to the AMA, there are a number of things to consider.
Physicians should be full partners at every stage of the AI lifecycle, the group said, from AI design to clinical integration, since they’re uniquely qualified to determine whether an AI tool is valid for a given indication.
There should also be a coordinated, transparent, whole-government approach, the group said. There needs to be clarity and consistency for developers, deployers and end users, including physicians and patients. Fragmented or duplicative rules would slow innovation and confuse clinicians; federal entities will have to act together to create a coherent oversight system.
