Perspectives on healthcare artificial intelligence policy from health equity professionals: findings from an...

While artificial intelligence tools (AI) can have positive impacts on healthcare, such as by expediting administrative tasks and aiding in clinical decision-making, recent studies have shown that AI-enabled systems can perpetuate biases against medically underserved populations which can exacerbate health disparities.

This qualitative interview study solicited the perspectives of professionals who work on issues related to health disparities and health equity to understand how health AI policy could be developed with considerations of addressing these issues. We conducted semi-structured interviews with individuals (n = 22) who work at organizations or serve in roles that focus on health disparities, promote health equity, and or primarily support health for underserved populations across the United States. The interview guide was developed using Michener's Racial Equity Framework for Assessing Health Policy with questions that focused on actors, networks, institutions, and contexts that are relevant to artificial intelligence, health equity, and health policy.

Three key themes emerged from our thematic analysis: 1) developing health AI policy for health equity starts with data, 2) health AI policy for health equity must include multiple institutions and strategies, and 3) considering economic issues is key for developing health AI policy that advances health equity. Participants highlighted the importance of interdisciplinary cross-sectoral collaboration to address the structural issues that contribute to health disparities as important for developing policy for health AI tools.

The findings of this study show that health equity professionals have specialized knowledge and offer unique insights important to emerging policy areas for health AI. This is the first study to focus on the views of this specific group on the development of health AI policy for health equity.