Co-Designing AI For Mental Health Support With Young Adults of Color (YOC): Needs, Expectations, and Implications for AI Literacy
This study addresses the cultural insensitivity, privacy concerns, and limited AI literacy prevalent in mental health AI systems serving youth of color. Employing a participatory co-design methodology, we conducted two-day workshops with 13 Asian, Black, and Latinx adolescents, complemented by qualitative interviews to investigate their needs and design preferences for AI chatbots. Our findings reveal how generic recommendations flatten individual lived experiences. We propose a novel framework integrating AI literacy with mental health literacy and reconceptualize the privacy calculus model. Furthermore, this work elucidates identity-based disclosure needs and expectations regarding AI accountability. Collectively, these insights offer critical guidance for developing user-centered, culturally responsive mental health AI systems that respect diverse sociocultural contexts.