Health-related artificial intelligence (health AI) systems are being rapidly created, largely without input from racially minoritized communities who experience persistent health inequities and stand to be negatively affected if these systems are poorly designed. Addressing this problematic trend, we critically review prior work focused on the participatory design of health AI innovations (participatory AI research), surfacing eight gaps in this work that inhibit racial health equity and provide strategies for addressing these gaps. Our strategies emphasize that “participation” in design must go beyond typical focus areas of data collection, annotation, and application co-design, to also include co-generating overarching health AI agendas and policies. Further, participatory AI methods must prioritize community-centered design that supports collaborative learning around health equity and AI, addresses root causes of inequity and AI stakeholder power dynamics, centers relationalism and emotion, supports flourishing, and facilitates longitudinal design. These strategies will help catalyze research that advances racial health equity.
https://dl.acm.org/doi/10.1145/3706598.3713165
The ACM CHI Conference on Human Factors in Computing Systems (https://chi2025.acm.org/)