# The Use of Large Language Models in Education **Author:** W. Xing, N. Nixon, S. Crossley and colleagues **Published:** 4 February 2025 **Length:** 8 mins (1,885 words) ## In Summary Introducing a special journal issue, Xing and colleagues survey emerging applications of large language models in education. They organise the contributions into three areas: automating tasks such as feedback analysis and assessment, adapting models to specific educational contexts, and supporting personalised learning environments. Reported results include promising performance in selected grading and feedback tasks, alongside mixed outcomes for AI-revised learning materials. The authors emphasise the need for rigorous evaluation of educational value, accuracy, bias and effects on teaching and learning. **Source:** [https://link.springer.com/article/10.1007/s40593-025-00457-x](https://link.springer.com/article/10.1007/s40593-025-00457-x) ## What might this mean for us? How might Higher Education distinguish successful automation from meaningful improvements in learning? Electric Dreams' assessment, feedback and dialogic documentation pilots offer opportunities to evaluate AI within particular disciplinary contexts. Could comparisons between automated outputs and academic judgement reveal where these systems support teaching, where they require human interpretation, and how their effects on students should be assessed? ## Links in this reading - [Recognizing multi-party epistemic dialogue acts during collaborative game-based learning, Acosta and colleagues](https://link.springer.com/article/10.1007/s40593-025-00457-x#ref-CR1) - [A large language model approach to educational survey feedback analysis, Parker and colleagues](https://link.springer.com/article/10.1007/s40593-025-00457-x#ref-CR16) - [Can LLMs grade open response reading comprehension questions?, Henkel and colleagues](https://link.springer.com/article/10.1007/s40593-025-00457-x#ref-CR7) - [Rewriting content with GPT-4 to support emerging readers in adaptive mathematics software, Norberg and colleagues](https://link.springer.com/article/10.1007/s40593-025-00457-x#ref-CR14) - [Automated scoring of constructed response items in math assessment, Morris and colleagues](https://link.springer.com/article/10.1007/s40593-025-00457-x#ref-CR12)