Abstract
The increasing use of artificial intelligence (AI) in healthcare has heightened public focus on issues of trust, transparency, and governance. In this study, we conducted five virtual community deliberations with Michigan residents (n = 159) to explore their hopes, concerns, and perspectives on how to promote trustworthiness in health AI. Participants were predominantly female (65%) with a mean age of 46 years, including 35% African American, 33% White, and 21% Middle Eastern/North African residents, with 40% reporting incomes below $50,000. Participants recognized AI's potential to improve diagnostic accuracy, efficiency, and access to health information, but voiced concerns about privacy, lack of transparency, diminished human interaction, and insufficient oversight. They recommended clear communication about when and how AI tools are used; regulation and oversight that includes patient and community input; and the use of AI to augment, not replace, human judgment and empathy. These findings informed the development of practice and policy recommendations for trustworthy health AI frameworks emphasizing human-centeredness, transparency, oversight, and accountability, underscoring the critical role of public involvement in AI governance. Future research should engage diverse community perspectives when developing health AI tools to support safe, trustworthy implementation in healthcare. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s43681-026-00987-7.