Autonomy versus algorithm: a replication study of student perspectives on AI ethical boundaries

Abstract

The widespread adoption of generative artificial intelligence (AI) tools in higher education has intensified the need to address their ethical implications. Building on a prior study of undergraduate perspectives on AI ethics, this research surveyed 200 undergraduates from multiple academic disciplinesat universities in Nigeria to explore their agreement with five AI ethics dimensionsbeneficence, non-maleficence, justice, autonomy, and explicabilityusing generative AI tools as the reference point. A mixed-method approach combined quantitative measures with qualitative explanations of participants’ views. Results indicated that autonomy received the highest agreement, while explicability was rated lowest, suggesting students were less concerned about the transparency of AI processes than about maintaining independent decision-making. Multiple regression analysis identified technology proficiency as a significant predictor of beneficence and gender as a predictor of non-maleficence, while academic level showed no significant influence. Thematic analysis revealed concerns about misinformation, erosion of creativity, job displacement, bias in AI outputs, and privacy risks, alongside counter-arguments emphasizing personal responsibility and trust in AI developers. These findings highlight the need for tailored ethical guidelines, targeted awareness initiatives, and integration of AI ethics into higher education curricula to foster informed and responsible AI use. By situating the research in the Nigerian higher education context, this study also considers how regional differences in educational systems, technology access, and cultural perspectives may shape attitudes toward AI ethics.

Ähnliche Arbeiten