JMIR Publications

ABSTRACT

Background:

Asian American emerging adults use mental health services at lower rates than the general US population. As generative artificial intelligence (AI) becomes a source of mental health support for young people, little is known about how Asian American emerging adults perceive these tools, decide whether to use them, and how they use them when they do.

Objective:

This study aimed to explore how Asian American emerging adults perceive AI for mental health, under what conditions they accept, use, or reject it, and how they utilized AI for their mental health.

Methods:

We conducted semi-structured interviews by Zoom with 38 Asian American emerging adults (aged 19-30 years; mean 25.2, SD 3.2) between August 2025 and December 2025. Participants were recruited from a parent survey examining mental health and help-seeking among Asian American emerging adults. The interview guide included dedicated questions about AI. We analyzed 177 interview excerpts about AI, drawn from 37 participants, using reflexive thematic analysis. Our team consisted of eight coders working in pairs, with the first author making final decisions on the analysis.

Results:

We constructed three themes. In Theme 1, “A More Personalized Google,” participants used AI for specific, bounded problems, such as a conflict with a coworker or deciding what to do first when overwhelmed, and valued structured, actionable recommendations over open-ended emotional processing. Participants used relational language such as “therapist” or “friend,” yet the functions they described remained focused on practical guidance. In Theme 2, “It’s Really Accessible,” participants turned to AI when professional care or informal support was delayed, effortful, unavailable, or socially costly. In Theme 3, “Not for Mental Health,” participants set boundaries based on the absence of human presence and empathy, concerns about sycophancy and harm, and values-based objections, and some declined to use AI at all.

Conclusions:

Participants described valuing AI’s practical guidance and accessibility while setting boundaries around its use. Their emphasis on concrete recommendations is consistent with counseling research on Asian American clients, which suggests that structured support may fit some users’ preferences across sources of help. Because this study had no comparison group, we cannot say whether the pattern is specific to Asian American emerging adults. These findings support evaluating coaching-oriented AI that accommodates individual preferences and keeps pathways to culturally responsive human care open.