A Comparative Study of Human and AI-Assisted Assessment using ChatGPT: The Case of Moroccan EFL Learners

Artificial intelligence (AI) is capable of performing various tasks and has been applied across multiple fields, including higher education. AI is increasingly used in teaching, learning, and classroom assessment. The primary aim of this study is to compare human assessments and AI-assisted assessments in Moroccan higher education, focusing on summative e-assessments. The primary question of the study is: How does AI-assisted assessment differ from human assessment in evaluating assignments in Moroccan higher education? The study yields data specific to the Moroccan context, filling the gap in this concern; it also provides educators and policymakers with a clear view of the similarities and differences between human and AI-assisted assessment. Writing assignments from EFL classes in two departments at the Faculty of Letters and Humanities, Dhar El Mahraz, were graded by their respective instructors and ChatGPT-4. The results of this study showed that AI grading produced higher scores than human grading, with differences that are both statistically significant in the department of French language studies, while in the department of psychology, the differences between human grading and AI grading are neither statistically significant nor practically large, suggesting that AI grading is comparable to human grading. The study’s findings imply that the grading system for AI might be different and is, therefore, incapable of replacing the human aspect of the grading process in Moroccan higher education, opening the possibility to enhance AI models of assessing EFL and align them with human standards of assessing EFL.