Psychological First Aid by AI: Proof‐of‐Concept and Comparative Performance of ChatGPT‐4 and Gemini in Different Disaster Scenarios
Journal of Clinical Psychology · 2025
DOI: 10.1002/jclp.23808 ↗Kurz gefasst
Vergleichsstudie in der Zeitschrift Journal of Clinical Psychology, 2025. Verfasst von Jun Tat Tan und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Old Age Psychiatry, Royal Cornhill Hospital, Aberdeen, UK. Hauptthemen laut MEDLINE: Disasters, Artificial Intelligence, First Aid, Hallucinations. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: This research was partially funded by the University of Arizona, Eller McGuire research grant (5642634/SMGNT)..
Abstract (Originalsprache)
Objective
This study aimed to evaluate the performance and proof‐of‐concept of psychological first aid (PFA) provided by two AI chatbots, ChatGPT‐4 and Gemini.
Methods
A mixed‐method cross‐sectional analysis was conducted using validated PFA scenarios from the Institute for Disaster Mental Health. Five scenarios representing different disaster contexts were selected. Data were collected by prompting both chatbots to perform PFA based on these scenarios. Quantitative performance was assessed using the PFA principles of Look, Listen, and Link, with scores assigned using IFRC's PFA scoring template. Qualitative analysis involved content analysis for AI hallucination, coding responses, and thematic analysis to identify key subthemes and themes.
Results
ChatGPT‐4 outperformed Gemini, achieving an overall score of 90% (CI: 86%–93%) compared to Gemini's 73% (CI: 67%–79%), a statistically significant difference ( p = 0.01). In the Look domain, ChatGPT‐4 scored higher ( p = 0.02), while both performed equally in the Listen and Link domain. The content analysis of AI hallucinations reveals that ChatGPT‐4 has a relative frequency of 18.4% (CI: 12%–25%), while Gemini exhibits a relative frequency of 50.0% (CI: 26.6%–71.3%), ( p < 0.01). Five themes emerged from the qualitative analysis: Look, Listen, Link, Professionalism, Mental Health, and Psychosocial Support.
Conclusion
ChatGPT‐4 demonstrated superior performance in providing PFA compared to Gemini. While AI chatbots show potential as supportive tools for PFA providers, concerns regarding AI hallucinations highlight the need for cautious implementation. Further research is necessary to enhance the reliability and safety of AI‐assisted PFA, particularly by eliminating hallucinations, and to integrate the current advances in voice‐based chatbot functionality.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Old Age Psychiatry, Royal Cornhill Hospital, Aberdeen, UK.
- Unit for Research in Emergency and Disaster, Area of Preventive Medicine and Public Health, Department of Medicine, Universidad de Oviedo, Oviedo, Asturias, Spain.
- McGuire Center for Entrepreneurship, University of Arizona, Tucson, Arizona, USA.
- UAI Business School, Universidad Adolfo Ibáñez, Peñalolén, Chile.
- Medicure KB Clinic, Kota Bharu, Kelantan, Malaysia.
ORCID-Profile: Rick Kye Gan
Förderung
- This research was partially funded by the University of Arizona, Eller McGuire research grant (5642634/SMGNT).
Zitieren
APA
Tan, J. T., Gan, R. K., Alsua, C., Peterson, M., Sales, R. Ú., Gan, A. Z., Cernuda Martínez, J. A., & González, P. A. (2025). Psychological First Aid by AI: Proof‐of‐Concept and Comparative Performance of ChatGPT‐4 and Gemini in Different Disaster Scenarios. Journal of Clinical Psychology. https://doi.org/10.1002/jclp.23808
Vancouver
Tan JT, Gan RK, Alsua C, Peterson M, Sales RÚ, Gan AZ, et al. Psychological First Aid by AI: Proof‐of‐Concept and Comparative Performance of ChatGPT‐4 and Gemini in Different Disaster Scenarios. Journal of Clinical Psychology. 2025. doi:10.1002/jclp.23808
BibTeX
@article{Tan2025,
author = {Tan, Jun Tat and Gan, Rick Kye and Alsua, Carlos and Peterson, Mark and Sales, Ricardo Úbeda and Gan, Ann Zee and Cernuda Martínez, José Antonio and González, Pedro Arcos},
title = {Psychological First Aid by AI: Proof‐of‐Concept and Comparative Performance of ChatGPT‐4 and Gemini in Different Disaster Scenarios},
journal = {Journal of Clinical Psychology},
year = {2025},
doi = {10.1002/jclp.23808}
}
EUVIMED