Navigating Artificial Intelligence in Assessment: Academic Professionals’ Perceptions of AI Grading in South Africa

Abstract

This study explores the perceptions of academic professionals in South Africa regarding the use of Artificial Intelligence (AI) in grading within higher education. In the current technological and digital era, the rapid integration of AI tools such as ChatGPT into academic environments points out that there is a growing need to understand how educators engage with these technologies, especially in assessment practices. This study adopted a mixed-methods explanatory design. The study first surveyed 56 academic professionals across diverse disciplines and then conducted semi-structured interviews with five participants to gain deeper qualitative insights. The study findings reveal a generally positive outlook toward AI-assisted grading, with many respondents acknowledging its potential to enhance efficiency, objectivity, and timely feedback. However, significant concerns emerged regarding the loss of qualitative judgment, contextual sensitivity, and the ethical implications of data use. The study highlights a disconnect between institutional support for AI adoption and the lack of structured training for academic staff. Participants expressed a clear preference for a hybrid grading model, combining AI tools with human oversight to preserve pedagogical integrity. Drawing on the Technology Acceptance Model (TAM), the research underscores the importance of perceived usefulness and ease of use in driving AI adoption, while also emphasising the socio-cultural and ethical contexts unique to South African higher education. The study concludes by recommending comprehensive AI training, robust ethical frameworks, and inclusive policy dialogues to ensure responsible, context-aware implementation of AI in assessment. These findings offer critical insights for institutions seeking to balance innovation with educational equity and integrity.