This analysis reveals improvements in feedback quality from generative AI systems in higher education, highlighting implications for pedagogical effectiveness.
The rapid development of Generative AI (GenAI) has opened new possibilities for its use in higher education, particularly in assessment and formative feedback. This study investigates the pedagogical effectiveness of GenAI-generated feedback using Systemic Functional Linguistics and Appraisal Theory to analyze the language used by GenAI reviewers. We compared two sets of GenAI-generated reviews on student writing from a graduate program in an American university. The first set came from a platform connected to OpenAI’s GPT-3, while the second used GPT-4, customized with a 35-million-word disciplinary corpus. The second version aimed to align more closely with the program’s academic context and provide more relevant, theoretically grounded feedback to the students enrolled in it. Through discourse analysis, we identified linguistic features that made the calibrated AI reviewer more pedagogically effective. Our findings highlight how tailoring GenAI systems to disciplinary language and feedback frameworks can improve the quality of support offered to university students. Based on our results, we also discuss pedagogical implications and offer recommendations for further research.
No takes yet. Share an insight, caveat, or question.
Zapata et al. (2025) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: