AI for Engagement and Motivation in Academia

central concern in academia is ensuring students to be motivated and engaged during work. Based on Ryan & Deci’s (1985) Self-Determination Theory, students feel engaged and motivated when three fundamental psychological needs, namely autonomy, competence and relatedness are fulfilled. Responding to increasing diffusion of digital affordances, this study investigates whether ChatGPT’s human like conversational and relational capabilities can fulfil students’ needs and ensure motivation and engagement. A between-subject experimental design, in which participants reported their perceptions of different manipulated versions of ChatGPT assisting with academic work. A total of 126 students were recruited and four roughly equal groups were questioned on four variations of the relational and conversational capabilities. The data was analysed through Hayes PROCESS macro (Model 4), to investigate the multiple effects of both direct and indirect paths of the research design. Results first ascertained the assumption of SDT, positioning the three psychological needs as prerequisites for students’ motivation and engagement. While the direct effects of ChatGPT’s capabilities were insignificant, significant mediation paths were confirmed. Both students perceived autonomy and competence positively affected students’ motivation and engagement. Though, this was not the case for perceived relatedness. Additionally, comparative analysis of the four conditional groups, while one heightened capability results in an increase in motivation and engagement, the combination of both sophisticated capabilities resulted in the strongest effect. Conclusively, ChatGPT’s capabilities provide a effectively support for students to work both motivated and engaged. However, this does not apply to a perceived feeling of relatedness. ChatGPT struggles with creating meaningful bonds and does not solve the common problem of feeling lonely and isolated during online work. AI tools, including ChatGPT, are pushing the boundaries of digital affordance capabilities, though they still lack effective mechanisms for interpersonal communication.