New ACP guidelines say AI should support, not replace, physician reasoning

In this video, Nadia Daneshwar, JD, MPH, ACP health IT policy specialists discussed the organization’s recent recommendations on the development and use of AI in health care.

“One of the main reasons physicians need these recommendations has to do with patient safety,” Daneshvar said. “The fact that these technologies are now embedded in various health systems and tools, and that physicians and other clinicians are using them without their knowledge, creates potential risks to patient safety.”

The ACP recommends:

• AI complements doctors’ reasoning and decision-making, not replaces it.
• The development, testing and use of AI must be in line with principles of medical ethics and decision-making.
• The development, testing and use of AI for patient care should be transparent.
• AI developers, researchers, and implementers must prioritize privacy and confidentiality of patient and clinician data.
• Clinical safety and effectiveness, as well as health equity, should be top priorities for AI developers, regulators, and implementers.
• AI should reduce healthcare disparities, not exacerbate them.
• AI developers should take responsibility for their models.
• AI tools should be designed to reduce the burden on clinicians at every stage of their development.
• Training in the use of AI should be provided at all levels of medical education.
• The environmental impact of AI and ways to mitigate it need to be studied and considered.

Dhaneshwar noted that the ACP's recommendations differ from those of other organizations for a few reasons: “We discuss the importance of not using anthropomorphic language in relation to AI technology and the impact that can have,” she explained.

reference: