University students face growing mental health challenges, yet continue to encounter barriers to in-person care, including access, stigma, cost, and limited capacity. As artificial intelligence (AI) becomes integrated into daily life, tools such as ChatGPT are increasingly explored by students for emotional support. This study examined how university students perceive the benefits, limitations, and potential utility of ChatGPT and other AI platforms as tools for mental health support. A cross-sectional survey was conducted among 392 students at the University of British Columbia, in which 259 (66.6%) had used ChatGPT for mental health. Quantitative data was analyzed using descriptive statistics, and qualitative responses underwent thematic exploration. Most students viewed ChatGPT as somewhat trustworthy and helpful for mental health-related questions, valuing its accessibility and non-judgmental space for emotional expression. Compared with professional support, 52.1% perceived AI as similarly useful, while 23.6% perceived it as better suited in certain situations, such as moments of hesitation to seek help. Non-users most frequently cited privacy concerns, inaccurate responses, and lack of human empathy as reasons for avoidance. Key limitations included reduced likelihood to seek professional help and limited human-like support. Overall, findings suggest that AI tools, such as ChatGPT, may serve as low-barrier emotional outlets and resources that complement counseling. Integration within academic settings may help bridge unmet needs, provided that ethical and privacy safeguards are prioritized. Further longitudinal research is warranted to clarify ChatGPT’s evolving role in digital mental health and its long-term impact on student wellbeing.