Abstract:
In Ukraine, university students face ongoing stress and educational disruption due to wartime conditions. In response to these issues, tools have been developed to support mental health as part of the Mental Health and Psychosocial Support System. Information about these resources is widely shared within the university environment. These tools include mental health chatbots, academic assistance applications, and personalized learning platforms, all utilizing artificial intelligence (AI) technologies. It is essential to study the effectiveness of these tools from the users' perspective as psychological assistance continues to evolve. This study aims to assess the impact of AI technologies on student well-being. Method. Data was collected from April 18, 2025, to May 8, 2025. The survey was part of a broader project studying well-being during wartime. One key question addressed the effectiveness of AI-based tools in coping with stress. Respondents could choose: "Helped", "Not helped", " Undecided", "Didn’t use". Participants (N=456; 69.3% female) from various Ukrainian universities completed the survey after informed consent. The majority were medical students (42.8%), followed by future teachers (25.7%) and psychologists (7.0%). Mental health was assessed using PHQ-9 and GAD-7 (cutoff: 10 points). Descriptive statistics, cross-tabulation, and ANOVA were conducted using R Studio. Results. Responses did not differ by gender (p>0.4). Overall, 56.6% (n=258) reported using AI, 34.2% (n=156) did not, and 9.2% (n=42) left the question blank while answering other sections. Among AI users, 39.5% (n=102) found it helpful, 41.1% (n=106) said it was not, and 19.4% (n=50) were undecided. No significant differences were found by specialty (p>0.5) or age (p>0.3). Anxiety and depression scores were significantly higher in females (p<0.001), with no link to specialty (p>0.1). Depressive symptoms were present in 39.6% of females and 21.4% of males; anxiety symptoms were found in 27.2% of females and 7.9% of males.PHQ-9 scores ranged from 8.1 (SD=6.1) to 10.1 (SD=5.6) with no significant differences by AI use (F=1.0; p=0.408). GAD-7 scores varied significantly (F=2.9; p=0.021): Helped by AI: 7.3 (SD=2.1); Not helped: 7.9 (SD=5.0); Undecided: 7.8 (SD=4.7); Didn’t use AI: 6.1 (SD=4.3). Participants who skipped the AI question in the questionnaire scored an average of 7.6 (SD=4.8) points. Conclusions. AI tools are commonly used as coping aids, but their perceived effectiveness is mixed. Students who used AI did not report better mental health than those who did not. The highest anxiety scores were found among students who avoided answering the AI question despite completing the rest of the survey. This group may represent a vulnerable subgroup that avoids digital mental health tools due to stigmatization, distrust, or privacy concerns. These findings suggest that AI use may reflect need rather than effectiveness and highlight the importance of addressing stigma in mental health support. Further research is needed to explore these issues.
