F1000Res. 2025 Oct 6;14:665. doi: 10.12688/f1000research.166372.2. eCollection 2025.
ABSTRACT
BACKGROUND: Barriers to mental-health care include limited resources and workforce, access constraints, and stigma. Artificial-intelligence (AI)-enabled chatbots may offer low-threshold support.
METHODS: Cross-sectional correlational (comparative) study in one private health-sciences university in the UAE. Proportional stratified random sampling across four colleges yielded n = 298 undergraduates. Instruments: (i) Socio-Demographic Questionnaire; (ii) researcher-developed AI Chatbot Usability questionnaire (content validated by a bilingual specialist); and (iii) Depression Anxiety Stress Scale-21 (DASS-21; established reliability/validity). Questionnaires were administered face-to-face with standardized instructions.
RESULTS: 206/298 (69.1%) had ever used an AI chatbot; most used Snapchat AI (76.9%), followed by ChatGPT/Bard (23.4% each). Overall, 57.0% had moderate-to-extremely-severe depression, 68.5% anxiety, and 33.6% stress. Users had higher odds of moderate-to-extremely-severe anxiety and depression than non-users. In multivariable models, higher depression (OR = 1.022; 95% CI 1.01-1.085; p<0.001) and anxiety (OR = 1.05; 95% CI 1.01-1.21; p<0.001) independently predicted chatbot use; stress did not.
CONCLUSION: Among UAE health-sciences students, AI-chatbot use is common and associated with higher depression/anxiety severity; this likely reflects help-seeking rather than causation. Universities should integrate early, stigma-sensitive supports, potentially including regulated, evidence-based chatbot tools-within stepped-care services.
PMID:41070110 | PMC:PMC12504930 | DOI:10.12688/f1000research.166372.2
