Youth mental health is a global crisis. Yet demand for support exceeds service capacity. Concurrently, young people are turning to general-purpose conversational AI for social and emotional support. These systems, however, have not been evaluated against real-world, representative benchmarks that include vulnerable youth, despite their potential detrimental psychological and social effects. Health agencies therefore call for responsible, child-centred design, human oversight and AI literacy.

This interactive and interdisciplinary workshop addresses two questions: What AI literacy promotes safe use of conversational AI in mental health contexts? What technical risks should be identified, escalated and reviewed in practice? Collaboratively, participants will identify practical principles for AI literacy, risk assessment and responsible AI design to guide developers, practitioners and policymakers.

Programme

This 90-minutes interactive and interdisciplinary workshop seeks to identify key principles for AI literacy and responsible AI design within mental health. After the workshop, an interdisciplinary team will use the session outputs as a starting point for developing recommendations for guidelines for AI literacy in the context of Danish youth mental health.

Welcome and structure (5 min)

Presentations (25 min)

• “An overview of international guidelines for AI-literacy in Mental Health”
  by Katerina Drakos, PhD student, at Child and Adolescent Mental Health Center, Copenhagen University Hospital – Bispebjerg and Frederiksberg, Copenhagen, Denmark and Copenhagen Center for Social Data Science (SODAS), University of Copenhagen, Copenhagen, Denmark.
  
  
• “Safety in Mental Health Chatbots-Technical Considerations”
  by Ahcene Boubekki, PhD, and Line H. Clemmensen, Prof. of Mathematics and Computer Science, and DTU Compute Department of Mathematics, University of Copenhagen
  
  
• “Risks and Benefits of Conversational AI for Youth Mental Health Support”
  by Anne D Müller, PhD, Child and Adolescent Mental Health Center, Copenhagen University Hospital – Mental Health, Services CPH, Copenhagen, Denmark

–

Small-group discussions and online polling (15min)
(interdisciplinary) on key technical and psychological challenges in young people’s use of GenAI for mental health support

Small-group discussions and online summarization (15min)
(interdisciplinary) on key themes, solutions and recommendations on young people’s use of GenAI for mental health support.

Plenary discussion and online polling (25min)
on solutions and recommendations for AI literacy for youth mental health, prioritized by participants, to identify emerging areas of consensus, disagreement or uncertainty.

Closing and invitation for further collaboration (5min)

Level

Intermediate: Suitable for attendees with basic understanding or experience in psychology, behavioral science, human-computer interaction or computer science who want to solve societal challenges in an interdisciplinary environment.

Organizers

• Anne Dorothee Müller, Researcher, Child and Adolescent Mental Health Center – Research Unit, Copenhagen University Hospital, Capital Region in Denmark (Lead organizer)
• Stella Grasshof, Associate Professor, IT University of Copenhagen
• Eva Paraschou, Doctoral Candidate, Department of Applied Mathematics and Computer Science, Technical University of Denmark
• Line H. Clemmensen, Professor, DTU Compute Department of Mathematics, University of Copenhagen
• Katerina Drakos, Doctoral Candidate, Child and Adolescent Mental Health Center, Copenhagen University Hospital – Bispebjerg and Frederiksberg, Copenhagen, Denmark and Department of Psychology, Copenhagen Center for Social Data Science (SODAS), University of Copenhagen, Copenhagen, Denmark 
• Nicole Lønfeldt, Associate Professor and senior researcher at Child and Adolescent Mental Health Center, Copenhagen University Hospital – Bispebjerg and Frederiksberg, Copenhagen, Denmark and Department of Psychology, Copenhagen Center for Social Data Science (SODAS), University of Copenhagen, Copenhagen, Denmark.