יעל פלדמן מגור

Educators play a central role in higher education assessment practices. As artificial intelligence (AI) becomes increasingly integrated into these practices, it is essential to understand educators’ perceived fairness in AI-mediated assessment. Most research on AI-mediated assessment focuses on performance and efficiency, leaving ethical issues imbued in AI out of this discussion. This study examines educators’ perceptions of fairness in AI-mediated assessment in Swedish higher education. It has two main objectives: first, to develop and validate a reliable instrument to measure educators’ perceived fairness in the use of AI-assessment practices; and second, to explore educators’ perceptions of fairness in AI-enabled assessment of student work. Drawing on Organizational Justice theory, the study applies four dimensions of fairness: procedural, distributive, informational, and interpersonal, to the context of AI-mediated assessment in higher education. Using a mixed-methods design, we used confirmatory factor analysis to validate a four-factor model using Likert-scale survey items and thematic analysis of responses to two open-ended questions. The findings suggest that educators perceived informational fairness most positively, followed by interpersonal fairness. Procedural and distributive fairness were assessed more cautiously, with median scores at the midpoint of the scale. Participants also emphasized the importance of maintaining active human oversight of both assessment procedures and outcomes. The study provides a validated instrument for future research on perceived fairness in AI-assisted assessment, both for practical applications and for future research. Furthermore, it highlights the critical role of educators’ professional development in promoting fairness awareness and supporting the responsible use of AI in student assessment.