The rapid proliferation of large language model AI systems including ChatGPT, Claude, Gemini, and local variants had created an unprecedented phenomenon in the lives of students and young people: for the first time in history, a form of intelligent, responsive, non-judgmental, available, and free conversational presence is accessible to anyone with a smartphone and data connection. Young people, particularly those in contexts where mental health support was stigmatized, counselling services are absent, and trusted adult relationships are unavailable, are increasingly turning to AI systems for emotional support, psychological guidance, health advice, and personal mentorship. This qualitative study examined the phenomenon of AI trust among students in Ugandan secondary and tertiary institutions, exploring why students were seeking emotional and health support from AI, what they experience in these interactions, how these interactions affected their wellbeing and decision-making, and what risks and opportunities the phenomenon presents. Through 52 in-depth interviews and 8 focus group discussions with secondary and university students, counsellors, teachers, and parents, the study documents a complex, emotionally rich, and ethically challenging landscape. Findings revealed that students were turning to AI for counselling and health support because human alternatives were unavailable, stigmatizing, or untrustworthy, and that many describe AI interactions as providing emotional validation, cognitive clarity, and practical guidance that they had previously been unable to access. Simultaneously, critical risks emerged including AI-reinforced health misinformation, inappropriate emotional dependency, privacy vulnerability, and the substitution of AI interaction for professional intervention in crisis situations. The study called for a nuanced, evidence-based institutional response that neither dismisses nor uncritically celebrates the role of AI in student wellbeing.