EUVIMEDEUVIMED entdecken
QuerschnittstudieOpen Access

Evaluating the Quality of Psychotherapy Conversational Agents: Framework Development and Cross-Sectional Study

Kunmi Sobowale iD, Daniel Kevin Humphrey iD

JMIR Formative Research · 02.07.2025 · 9 · e65605-e65605

DOI: 10.2196/65605 ↗

Kurz gefasst

Querschnittstudie in der Zeitschrift JMIR Formative Research, online erschienen am 02.07.2025. Verfasst von Kunmi Sobowale und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, 760 Westwood Plaza, Suite 48-241, Los Angeles, CA, 90024,…. Hauptthemen laut MEDLINE: Psychotherapy, Artificial Intelligence, Communication. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Despite potential risks, artificial intelligence–based chatbots that simulate psychotherapy are becoming more widely available and frequently used by the general public. A comprehensive way of evaluating the quality of these chatbots is needed.

Objective

To address this need, we developed the CAPE (Conversational Agent for Psychotherapy Evaluation) framework to aid clinicians, researchers, and lay users in assessing psychotherapy chatbot quality. We use the framework to evaluate and compare the quality of popular artificial intelligence psychotherapy chatbots on the OpenAI GPT store.

Methods

We identified 4 popular chatbots on OpenAI’s GPT store. Two reviewers independently applied the CAPE framework to these chatbots, using 2 fictional personas to simulate interactions. The modular framework has 8 sections, each yielding an independent quality subscore between 0 and 1. We used t tests and nonparametric Wilcoxon signed rank tests to examine pairwise differences in quality subscores between chatbots.

Results

Chatbots consistently scored highly on the sections of background information (subscores=0.83-1), conversational capabilities (subscores=0.83-1), therapeutic alliance, and boundaries (subscores=0.75-1), and accessibility (subscores=0.8-0.95). Scores were low for the therapeutic orientation (subscores=0) and monitoring and risk evaluation sections (subscores=0.67-0.75). Information on training data and knowledge base sections was not transparent (subscores=0). Except for the privacy and harm section (mean 0.017, SD 0.00; t 3 =∞; P

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

PsychotherapyArtificial IntelligenceCommunicationHumansCross-Sectional Studies

Stichwörter der Autor:innen

AIChatGPTaccessibilityartificial intelligencechatbotscliniciansconversational agentdigital healthevaluation frameworkevaluation studygenerative AIlarge language modelspsychotherapypsychotherapy chatbotsresearchers

Autor:innen und Institutionen

  • Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, 760 Westwood Plaza, Suite 48-241, Los Angeles, CA, 90024, United States, 1 310-794-7035, 1 925-281-3270.
  • Department of Psychology, College of Arts and Science, University of San Francisco, San Francisco, CA, United States.

ORCID-Profile: Kunmi Sobowale, Daniel Kevin Humphrey

Zitieren

APA

Sobowale, K., & Humphrey, D. K. (2025). Evaluating the Quality of Psychotherapy Conversational Agents: Framework Development and Cross-Sectional Study. JMIR Formative Research, 9, e65605-e65605. https://doi.org/10.2196/65605

Vancouver

Sobowale K, Humphrey DK. Evaluating the Quality of Psychotherapy Conversational Agents: Framework Development and Cross-Sectional Study. JMIR Formative Research. 2025;9:e65605-e65605. doi:10.2196/65605

BibTeX
@article{Sobowale2025,
  author = {Sobowale, Kunmi and Humphrey, Daniel Kevin},
  title = {Evaluating the Quality of Psychotherapy Conversational Agents: Framework Development and Cross-Sectional Study},
  journal = {JMIR Formative Research},
  year = {2025},
  volume = {9},
  pages = {e65605--e65605},
  doi = {10.2196/65605}
}
Kostenlos starten →