This study examines the intention to adopt AI-powered chatbots among higher education students in India, focusing on the roles of performance expectancy and effort expectancy, and the mediating effect of perceived trust in linking social influence, perceived intelligence, and perceived risk with adoption intention. A quantitative survey of 575 students was analysed using Structural Equation Modelling in AMOS and the PROCESS macro, with reliability and validity confirmed through composite reliability (CR), average variance extracted (AVE), discriminant validity, and satisfactory model fit indices (CFI = 0.926; RMSEA = 0.04). The findings reveal that performance expectancy (β = 0.4729, p < 0.001) and effort expectancy (β = 0.2283, p < 0.001) significantly influence chatbot adoption intention, while perceived trust plays a central mediating role between social influence (β = 0.0724), perceived intelligence (β = 0.0656), perceived risk (β = −0.0979), and adoption intention. The study extends TAM and UTAUT by incorporating perceived intelligence, risk, and trust, offering deeper insights into AI adoption in higher education, and suggests that institutions should enhance chatbot usability, intelligence, transparency, and faculty and peer endorsement to strengthen trust and mitigate risk perceptions