AMA urges health systems to act now on AI governance

Editor's Note

Artificial intelligence use in healthcare is accelerating, and the American Medical Association (AMA) is pressing health systems to establish clear governance policies before the technology outpaces oversight. Nearly 70% of physicians reported using artificial intelligence tools in 2024, a sharp rise from 38% in 2023, AMA News Wire August 18 reports. Enthusiasm is also growing, with 35% of physicians more excited than concerned about AI compared to 30% the year prior.

To support safe adoption, the AMA, in collaboration with Manatt Health, released its “Governance for Augmented Intelligence” (AI) toolkit. The eight-step framework guides organizations in implementing, managing and scaling AI. It emphasizes executive accountability, formation of a governance group, evaluation of existing policies, development of AI-specific rules, vendor assessment, oversight processes, and staff training. The toolkit also includes a model AI policy document that organizations can adapt to their own structures.

As detailed in the news article, the AMA urges organizations to define key terms such as generative AI and machine learning, specify both permitted and prohibited uses, and establish transparency standards for when clinicians or patients should be informed that AI is being applied. For instance, using AI for drafting marketing materials may be permitted, but entering patient health information into public tools should be strictly barred. Policies should also address risk assessment, regulatory compliance, data retention, and accountability. Training programs are another essential element, ensuring all staff understand the boundaries of acceptable use.

Plus, health systems should review and update related policies, including antidiscrimination, data security, patient safety reporting, and codes of ethics, to ensure alignment with new AI rules. Margaret Lozovatsky, MD, AMA’s chief medical information officer, highlighted that technology is advancing faster than organizations can implement it, making governance urgent.

The AMA consistently refers to AI as “augmented intelligence” to reinforce its role in supporting, not replacing, clinicians. Beyond guidance, the association continues to advocate for strong oversight, transparency in AI use, clear liability standards, data privacy protections, and careful scrutiny of payer reliance on automated decision-making.