Artificial intelligence as a support tool for providing feedback on coursework in higher education: A pilot study

The increasing use of artificial intelligence (AI) in higher education is reshaping feedback practices and raising questions about how AI-generated feedback compares with feedback provided by humans. This pilot study investigates the pedagogical potential of an AI assistant for providing feedback[1] on student´ academic writing in higher education. Drawing on a sociocultural perspective on scaffolding and framework distinguishing between feed back and feed forward, the study examines (i) how students experience AI-generated feedback compared with feedback provided by learning assistants, (ii) the extent to which the two feedback sources align with predefined assessment criteria, and (iii) their similarities and differences in the content of the provided feed back and feed forward. Data comprised students´ survey responses and content analysis of feedback provided by the AI assistant and learning assistants on the same type of academic writing assignment. The findings show that both feedback sources largely align with the predefined assessment criteria and provide feedback on students´ current performance as well as feed forward aimed at further development. However, they differ in how this support is enacted. The AI assistant tends to provide more explicit, criteria-focused, and action-oriented guidance, whereas learning assistants provide more contextualized and relationally grounded feedback. Students also expressed a preference for feedback from learning assistants on future assignments. Viewed through a sociocultural lens, these findings suggest that AI-generated and human feedback may represent distinct but complementary forms of pedagogical scaffolding. Rather than replacing human feedback, AI may therefore serve as a pedagogical resource that combines explicit guidance with dialogic and relational support. [1] In this study, feedback is used as an overarching concept encompassing three complementary functions: feed up (clarifying goals and exptectations), feed back (information about current performance), and feed forward (guidance for future improvement).