Trust, Bias, and Agency: A Critical Examination of Students’ Perceptions of AI in Academic Contexts

Abstract

The fast incorporation of artificial intelligence (AI) in higher education has sparked an increased interest in understanding how students view its function, dependability, and ethical implications. This research critically investigates students' perceptions of AI in academic settings, focusing on the interconnected characteristics of trust, bias, and agency. Data from 60 undergraduate students, who were selected using purposive sampling technique, were acquired using a qualitative study approach, which included semi-structured interviews and theme analysis. The findings show that students have conditional faith in AI technologies, valuing their efficiency and accessibility while being dubious about accuracy and fairness. Participants were aware of algorithmic bias, notably in terms of Western-centric linguistic and cultural representation, but lacked a thorough knowledge of the technical causes of such bias. Furthermore, students navigated a complicated sense of agency, combining empowerment from AI-assisted learning with worries about reliance, authorship, and academic integrity. Emotions such as excitement, anxiety, and guilt highlighted the moral complexity surrounding AI's employment in academic research. These findings emphasize the need for higher education institutions to create critical AI literacy frameworks, clear policies, and context-sensitive pedagogies that promote ethical, egalitarian, and human-centred interaction with AI technology. The work adds to current discussions about human-AI partnerships in education by revealing how trust, bias, and agency are dynamically intertwined in student’s lived experiences with AI-assisted learning.