Session
Stress-test FRED: A mental health conversational AI agent for struggling families
Organized by:
Thursday 8 October 11.00
Lead organizer: Nicole Lønfeldt, Associate Professor, Copenhagen University Hospital
FRED (Family Resilience with Empathy and Dialogue) is an AI-based parent training conversational agent, designed to provide real-time personalized guidance for parents of children experiencing difficult emotions and challenging behaviors. The conversational agent aims to support parents across different stages of help-seeking, addressing an unmet need for families on waitlists.
The session will be led by an interdisciplinary group of data scientists, clinicians and researchers in psychiatry, and engineers from industry. The audience will engage in a practical stress-testing activity with FRED, exploring potential vulnerabilities that are absent from static LLM benchmarks.
The session opens with brief introductory talks covering FRED, associated clinical and ethical risks, and red-teaming methodology. Participants will then take part in a 50-minute interactive red-teaming activity, concluding with a plenary session to share group findings and discuss key takeaways.
Note: the organizers kindly request that audience members bring a laptop for participating in the interactive red-teaming activity.
11:00 – 11:05
Welcome
Nicole Lønfeldt, Psychologist and Associate Professor
Child and Adolescent Mental Health Center, Copenhagen University Hospital /
Copenhagen Center for Social Data Science and Department of Psychology, University of Copenhagen
11:05 – 11:10
Introduction to FRED
Simge Genç, PhD Fellow in Digital Mental Health
Child and Adolescent Mental Health Center, Copenhagen University Hospital /
Copenhagen Center for Social Data Science, University of Copenhagen
11:10 – 11:15
Risks and red-teaming
Sid Ravi, AI Scientist and Researcher
Interhuman AI
11:15 – 11:20
Groups and workshop website
Line Clemmensen, Professor in AI and Machine Learning
Department of Mathematical Sciences, University of Copenhagen /
Department of Applied Mathematics and Computer Science, Technical University of Denmark
11:20 – 11:25
Red-teaming instructions
Eva Paraschou, MCSA PhD Fellow in AI Alignment in Mental Health
Department of Applied Mathematics and Computer Science, Technical University of Denmark
11:25 – 12:15
Red-teaming activity
All participants
12:15 – 12:25
Group findings
Camilla Uhre, Psychologist and Senior Researcher
Child and Adolescent Mental Health Center, Copenhagen University Hospital /
Center for Clinical Neuropsychology, Children and Adolescents, Rigshospitalet
12:25 – 12:30
Closing remarks
Nicole Lønfeldt, Psychologist and Associate Professor
Child and Adolescent Mental Health Center, Copenhagen University Hospital /
Copenhagen Center for Social Data Science and Department of Psychology, University of Copenhagen
Introductory: Suitable for beginners with little to no prior knowledge of the subject.