From AI adoption to underperformance? Investigating the impact of interactive AI tools on student outcomes in higher education

This study explores the paradoxical effects of interactive AI tools on university students’ academic performance. Although technologies such as chatbots and virtual assistants are designed to enhance learning through personalised experiences, inadequate integration into curricula may lead to adverse outcomes. Grounded in the Unified Theory of Acceptance and Use of Technology 2 (UTAUT-2), this research examines determinants of students’ behavioural intention to use AI tools and their subsequent impact on academic achievement. Data collected from 407 students across Spanish and French universities reveal that performance expectancy, social influence, hedonic motivation, price value, and habit significantly predict intention to use, whereas effort expectancy and facilitating conditions do not. Importantly, actual usage of these tools shows a statistically significant negative association with academic performance, suggesting that unstructured adoption can be counterproductive. These findings underscore the need for institutional strategies to ensure pedagogically sound integration of AI technologies. While students perceive AI as a means to improve outcomes, improper use may result in lower grades. The study offers critical implications for higher education, highlighting the necessity of balancing technological innovation with effective instructional design to maximise learning benefits.