New Publication in The Lancet Psychiatry: LLMs as Mental Health Providers
Sentio Marriage and Family Therapy Program in California President Tony Rousmaniere, in collaboration with Simon B. Goldberg and John Torous, has published a new article in The Lancet Psychiatry titled "Large language models as mental health providers."
The piece examines the rapid integration of Large Language Models (LLMs) into the mental health landscape and the resulting clinical, ethical, and regulatory challenges.
Key Considerations for the Field
The authors address the "quasi-clinical" role that unregulated AI currently plays for millions of users. The article highlights several critical areas for professional and regulatory attention:
The Scale of Unregulated Support: With therapy-related AI interactions now numbering in the millions monthly, LLMs are operating as a de facto frontline mental health resource without the oversight of professional ethical codes or mandated reporting requirements.
Methodological Heterogeneity: Current evaluation methods for AI effectiveness remain inconsistent. The authors advocate for standardized, clinically informed frameworks to assess AI safety and efficacy.
Opportunities for Clinical Insight: The paper notes that the unprecedented scale of LLM interaction data offers a new frontier for understanding help-seeking behavior and treatment trajectories, provided rigorous safety and privacy protocols are established.
This publication represents an effort to bridge the gap between rapid technological advancement and the fundamental requirements of clinical safety and professional responsibility.
Access the Full Article: Rousmaniere, T., Goldberg, S. B., & Torous, J. (2025). Large language models as mental health providers. The Lancet Psychiatry.
For more information on the empirical research cited in this piece, visit the Sentio Marriage and Family Therapy Program in California research portal.
About this study
- Full title
- Large language models as mental health providers
- Authors
- Tony Rousmaniere (Sentio University), Simon B. Goldberg, and John Torous
- Publication
- The Lancet Psychiatry, 13(1), 7–9, January 2026 (Elsevier).
- Publication type
- Commentary (Comment section)
- DOI
- https://doi.org/10.1016/S2215-0366(25)00269-X
- Sentio's role
- Written by Sentio University President Tony Rousmaniere as lead author, with Simon B. Goldberg and John Torous.
- Main finding
- Large language models already operate as a de facto frontline mental health resource at a scale of millions of interactions per month, outside professional ethical codes, mandated reporting, and standardized evaluation, and the authors call for clinically informed frameworks to assess their safety and efficacy.
- How to cite
- Rousmaniere, T., Goldberg, S. B., & Torous, J. (2026). Large language models as mental health providers. The Lancet Psychiatry, 13(1), 7–9. https://doi.org/10.1016/S2215-0366(25)00269-X
This publication is part of the Sentio University AI research program. Sentio also offers an AI certification course for therapists and a Master of Arts in Marriage and Family Therapy built on Deliberate Practice.
