(PS8-59) Cross-cultural Perspectives on Ai-based Psychotherapy: An International Survey on Acceptance and Ethical Considerations

Abstract Text: Background
In recent years, with the advancement of large-scale language models and generative AI, attention has increasingly focused on their potential application to psychotherapy. Conversely, side effects—such as dependency, data privacy concerns, and disparities—have begun to be highlighted in the context of applying AI to psychotherapy (Jurblum, M. et al., 2025). However, few studies have specifically addressed these potential risks, and the influence of cultural factors, which have been shown to significantly affect attitudes toward AI, remains largely unexplored. Furthermore, no previous research has compared the needs for AI-based psychotherapy with those for face-to-face psychotherapy, remote psychotherapy, or app-based psychotherapy. Consequently, the Ethical, Legal, and Social Issues (ELSI) associated with AI-based psychotherapy remain unclear.
Against this backdrop, the present study conducted an international survey of the general public in Japan, the United States, and the United Kingdom. The aim was to elucidate how AI-based psychotherapy is perceived and accepted relative to face-to-face psychotherapy, remote psychotherapy, and app-based psychotherapy. .
Participants
From June 24 to July 4, 2024, participants were recruited via an internet survey company in Japan, the United States, and the United Kingdom. To prevent sampling bias, we ensured an even distribution by age (20s–80s) and gender (male/female). A total of 490 valid responses were collected from each country, yielding 1,470 respondents overall.
Survey Items
Attitudes toward AI psychotherapy were assessed using a questionnaire adapted from the international survey on new medical technologies by Shineha et al. (2022).The survey covered seven categories:
(i) Awareness, (ii) Development Duration, (iii) Opinion for Proceeding, (iv) Factors of Acceptance, (v) Intersting contents, (vi) Predict of Advancement, and (vii) Effectiveness and accessibility compared to face-to-face psychotherapy.The questionnaire included 65 items with response options ranging from 2 (e.g., Yes/No) to 13 (e.g., Risk, Cost, Mechanism, Safety Measures, Policy).
Data Analysis
Chi-square tests were conducted to examine differences in attitudes based on demographic variables. Standardized residuals were also calculated.
Results and Discussion
The results indicated that although fewer individuals were aware of AI-based psychotherapy compared to face-to-face psychotherapy, a notable proportion believed that AI-based psychotherapy should be further promoted. By contrast, when compared to face-to-face or other treatment modalities, security factors emerged as critical for the acceptance of AI-based psychotherapy. Additionally, both nationality and prior experience with psychotherapy were identified as factors influencing acceptance. In light of these findings, we are currently developing and evaluating the safety of psychotherapeutic interventions provided by AI agents employing generative AI. The outcomes of these studies will be presented in a forthcoming poster.