Co-Designing AI For Mental Health Support With Young Adults of Color (YOC): Needs, Expectations, and Implications for AI Literacy

Generic AI advice fails to account for lived experiences: Participants found that current mental health chatbots provide one-size-fits-all guidance that doesn't reflect the specific challenges young adults of color face, suggesting that effective AI mental health tools must move beyond standardized responses to incorporate identity-informed understanding.

Privacy concerns are intertwined with AI literacy gaps: The study reveals that young adults of color lack sufficient understanding of how AI systems work and handle their data, creating a critical need to integrate AI literacy education alongside mental health literacy to help users make informed decisions about sharing personal information with chatbots.

Young adults of color want clearer boundaries and accountability: Participants expressed confusion about whether they were receiving clinical or non-clinical support from chatbots and wanted greater transparency about AI limitations and stronger mechanisms for holding these systems accountable, indicating that current tools lack sufficient safeguards and clarity.

Identity-informed personalization is essential but requires careful design: Rather than making incorrect assumptions about users' backgrounds, participants preferred chatbots that allowed them to disclose their identity and cultural context, suggesting that responsible AI design must actively invite and incorporate this information rather than relying on demographic stereotyping.

Young adults of color should be centered in AI design decisions: The co-design approach itself was valuable, demonstrating that this population's perspectives are often missing from AI development and that their direct input is necessary to create mental health tools that are both effective and trustworthy.