Exploring Opportunities to Augment Psychotherapy with Language Models

Natural language data, like patient narratives, are crucial in psychotherapy, yet psychotherapists face challenges using these qualitative data to tailor treatment to patient needs. Innovations in natural language processing, including breakthroughs in language models (LMs), show opportunities like summarizing conversational data into quantitative information. In this study, we investigated how LM-based tools can augment patient measurements and treatment delivery in psychotherapy.

Through formative interviews and design provocation sessions with a total of six psychotherapists, we identified three opportunities: 1) to quantify and summarize extensive qualitative data for easier retrieval and monitoring of treatment progress; 2) to give clinicians a structured tool to support their patients’ learning; and 3) to facilitate treatment personalization. Our findings suggest that LM-based tools can potentially facilitate data-driven clinical practice, reduce cognitive and administrative burdens, and improve treatment quality. Additionally, our research paves the way for developing LM-based tools that can be integrated into psychotherapy.

Yang, Y.,* Viranda, T.,* Van Meter, A. R., Choudhury, T., & Adler, D. A. (2024). Exploring opportunities to augment psychotherapy with language models. In Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (CHI EA '24) (Article 144, pp. 1–8). Association for Computing Machinery, New York, NY, USA. https://doi.org/10.1145/3613905.3650990