Chatbot-Based Interventions for Mental Health Support

Abstract

Objective: Mental health concerns are rising, particularly among post-secondary students, who may lack access to traditional therapeutic resources due to barriers like long wait times and high costs. To help address these challenges, we explored the potential of large language model-based chatbots for supporting mental health and wellbeing in student populations.Methods: We conducted two studies, lasting one week and four weeks, to examine the effectiveness of chatbot interventions over different durations. Both studies compared two chatbot interventions—one mindfulness-focused and one value-focused—against an active check-in-only control condition. The primary outcome measure was the improvement in wellbeing through the mindfulness-to-meaning (MM) pathway, a process in which enhanced decentering, the ability to see one’s experience from a wider perspective, leads to improved positive reappraisal, the ability to find constructive and empowered interpretations of experience.Results: All conditions showed evidence of stress reduction. However, compared to the active control group, both intervention styles at both durations resulted in improved wellbeing via the MM pathway. This effect was primarily driven by significant improvements in decentering. For the longer duration only, we also observed enhanced reappraisal. Conclusions: These results emphasize the potential of chatbot-based interventions to support the development of regulatory skills by leveraging the MM pathway to enhance mental health. Educational institutions and mental health providers might consider integrating such tools into scalable and accessible student support systems, addressing a broader and more diverse audience while promoting sustained wellbeing through skills development.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 3 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.