From Hidden Barriers to Engagement: Workshop on AI-Supported Digital Storytelling for Immigrant Mental Health

Background: Immigrant communities face persistent barriers to mental health care, including stigma, language discordance, and limited culturally responsive services. Korean immigrant and Korean American adults experience elevated suicide risk and low utilization of formal support, while artificial intelligence (AI)–based mental health tools are largely designed for dominant-language users, often overlooking immigrant family dynamics, bicultural stress, and culturally specific help-seeking norms. These gaps raise critical concerns regarding equity, access, and representation in emerging digital mental health interventions.

Methods: This workshop introduces an AI-supported digital storytelling approach grounded in a human-centered AI framework and the data justice perspective proposed by Linnet Taylor (2017). Findings are drawn from a pilot study with Korean immigrants and Korean American adults (N = 15). Participants engaged in structured AI storytelling interactions followed by feedback collection. Descriptive frequencies were used to summarize user-reported experiences, and qualitative insights were integrated to contextualize engagement patterns. The workshop is designed as an interactive session integrating (1) conceptual foundations (human-centered AI, digital storytelling, and data justice dimensions of visibility, engagement, and cultural representation), (2) an experiential demonstration of AI-supported storytelling, (3) critical evaluation of limitations, and (4) applied discussion on integrating AI tools into culturally responsive social work practice.

Results: Findings indicate that AI-supported digital storytelling functions as a low-threshold entry point for mental health engagement. Over half of participants (53%, 8/15) described the tool as readily accessible and immediately usable, while 27% (4/15) emphasized its ability to support real-time emotional release. Notably, 20% (3/15) disclosed thoughts they could not share with a human counselor, suggesting reduced stigma and increased psychological safety. Additional benefits included enhanced reflection on anxiety (20%, 3/15), positive engagement with identity-focused activities such as motto creation (13%, 2/15), and overall satisfaction with usability and design (20%, 3/15). A subset of participants (13%, 2/15) specifically highlighted the importance of interacting in Korean, underscoring the role of language concordance in engagement.

At the same time, substantial limitations emerged. A majority of participants (80%, 12/15) reported contextual inconsistency, noting that the AI failed to maintain conversational continuity or generated irrelevant responses. Similarly, 67% (10/15) described interactions as repetitive or fragmented, indicating limitations in sustained narrative understanding. Nearly half (47%, 7/15) reported a lack of emotional empathy compared to human counselors. Additional concerns included unclear or unwelcoming interface design (47%, 7/15), technical barriers (27%, 4/15), and loss of cultural nuance (27%, 4/15).

Conclusions & Implications: AI-supported digital storytelling offers an accessible, culturally adaptable entry point for mental health engagement among immigrant populations, but shows limitations in emotional attunement, contextual understanding, and cultural nuance. Applying a data justice lens, this workshop emphasizes balancing disclosure, narrative expression, and (anti)discrimination (reducing bias and misrepresentation). AI should be positioned as a complementary tool that supports early engagement while maintaining human oversight and culturally responsive design. These findings underscore the role of social work in guiding the ethical and equitable integration of AI to reduce barriers and expand meaningful access to care.