A new commentary written by physicians in the American Journal of Public Health concludes that AI, in its current state, cannot be ethically or responsibly implemented in public health systems. CIDRAP has the news.

The physicians said that AI poses risks to patients, especially those from historically marginalized groups. They warn that the technologies are being implemented haphazardly and without adequate vetting into public health. They caution that many of the tools are "being deployed amid fragmented regulatory frameworks, validation standards, and equity safeguards that govern other health interventions."

Data used to train AI could be reflective of historical trends of poor healthcare access, structural racism, and social inequality. Generative AI could produce intentionally or unintentionally false health information that can mislead people into making poor decisions. These tools could also provide uniform health recommendations (i.e. annual mammograms for all women aged 40 and older) despite ongoing debates about the efficacy of certain practices in the medical community. This could especially be harmful to people with limited health literacy or English proficiency.

The physicians suggested several strategies to assess AI tools, including mandatory equity-impact evaluations; validations in the settings for which they are intended; alignment with data-sovereignty principles; and transparency regarding how and when AI is used in communities.