Presentation A Piloting Study for a Human-Computer Integrated Feedback Approach in Academic Writing more

Providing feedback on student writing in a second language is among teachers’ most vital but demanding responsibilities. Automated Writing Evaluation (AWE) tools, like Grammarly and ChatGPT, have surfaced as promising aids to assist teachers in this task, yet the balance and collaboration between human and machine-generated feedback seem to remain a dynamic and evolving aspect as technological capabilities progress. Therefore, this study aims to investigate the differences in the types of feedback provided by ChatGPT versus the integrated feedback from Grammarly and teachers. Thus, this research seeks to explore the variations in feedback types generated by ChatGPT compared to the combined feedback from Grammarly and a lecturer at a state university. The five participants were randomly-chosen first-year students studying in an English Language Teaching Program. The results revealed variations in feedback distribution depending on the individual student’s writing, meaning some essays triggered more AI feedback while others required more human intervention. Teachers and Grammarly seem to focus on grammar and structure more, while ChatGPT might emphasize clarity, cohesion, and argumentation. This study will explore the reasons behind this variation and the feedback types emphasized, aiming to propose a hybrid approach to feedback in academic writing.