Using AI to Co-Create Digital Mental Health Tools with Teens in Rural South Africa: A Qualitative Study
Background:
Digital mental health interventions (DMHIs) offer a scalable solution to address adolescent depression and anxiety. User-centered co-production can enhance acceptability and engagement, but it's often resource-intensive.
Advances in generative AI (GenAI) present new opportunities for involving teens in co-design, yet research on its feasibility and acceptability, especially in low-resource settings, is limited.
Objective:
This study aimed to explore teens' experiences and perspectives of using GenAI to co-create stories, images, and music for the Kuamsha app, a gamified DMHI teaching behavioral activation through interactive narratives and peer support.
Methods:
Two participatory workshops and focus group discussions were conducted with 23 teens (aged 15-19) in rural South Africa. Participants used 3 GenAI tools (ChatGPT, MidJourney, and Soundful) to create digital content. Data were audio-recorded, translated, transcribed, and triangulated with facilitator notes. Thematic analysis was used to explore key themes.
Results:
Almost all participants (22/23, 96%) had no prior GenAI exposure. Most (20/23, 87%) described the creative process as enjoyable and engaging, with many (21/23, 91%) reporting improved mood from creating music. Teens expressed autonomy and ownership, with over half (14/23, 61%) personalizing outputs to reflect their identities and aspirations. All participants (23/23, 100%) preferred AI-generated images over the cartoon-like illustrations in Kuamsha, and most (19/23, 83%) preferred AI-generated music. Story preferences were mixed, with a quarter (6/23, 26%) noting that Kuamsha's narratives contained embedded lessons not integrated into ChatGPT outputs.
Most teens (18/23, 78%) needed support with prompt construction, and over half (13/23, 57%) noted cultural biases in AI outputs, especially in images. Most participants (17/23, 74%) expressed interest in using AI for schoolwork and creative projects, while a minority (6/23, 26%) preferred to limit use to personal applications. Concerns about fairness and the displacement of human creativity were also raised.
Conclusions:
GenAI shows promise for enhancing teen engagement in DMHI co-production and enabling culturally relevant, personalized content. However, reliance on human support and persistent algorithmic biases remain limitations. Further research should explore integrating therapeutic principles into AI-generated media and strategies to mitigate bias.
Controversy:
GenAI's potential to revolutionize DMHI development is exciting, but its ethical implications are complex. While GenAI can create culturally relevant content, it also risks perpetuating biases and stereotypes. How can we ensure that GenAI tools are inclusive and empowering, especially for underrepresented groups?
Comment:
This study highlights the importance of user-centered design in DMHI development. By involving teens in the co-creation process, we can create more engaging and effective treatments. However, the need for human support and the presence of algorithmic biases are challenges that require further exploration. What strategies can we employ to address these issues and ensure that GenAI-assisted co-design is both effective and ethical?
