Affordances and Constraints of AI-Generated Writing Feedback

When AI-Mediated Feedback Writing Works and When It Does Not: Lessons Learned from Two Classroom Cases

EFL Students’ Perceptions on AI-Generated Feedback in Academic Writing

The increasing use of Artificial Intelligence (AI) writing tools has changed the way English as a Foreign Language (EFL) students seek feedback on their academic writing, particularly during the demanding process of preparing a research proposal or thesis. This study aims to explore how final-semester EFL students perceive, evaluate, and respond to AI-generated feedback while writing their proposal or thesis. Using a descriptive qualitative design, this study employed a Likert-scale screening questionnaire distributed to 24 respondents, followed by semi-structured interviews with six key informants selected through purposive sampling. The findings of the questionnaire indicate that most participants perceived AI-generated feedback favorably for surface-level writing issues, with 87.5% agreeing that it helped them identify possible errors before consulting their supervisor, while only 62.5% reported cross-checking content-level AI suggestions before applying them. From a qualitative perspective, thematic analysis organized these perceptions into four components: cognitive (understanding of how AI feedback is produced and how it differs from supervisor feedback), affective (confidence balanced with doubt about accuracy), conative (direct application of surface-level suggestions versus cross-checking of content-level suggestions), and external/internal cues (trust shaped by feedback consistency and by students’ own digital literacy). When examined through Rosenberg and Hovland’s Tripartite Attitude Model, Robbins and Judge’s theory of perception, and the concept of AI Literacy, the findings show that students who engaged more critically and selectively with AI-generated feedback demonstrated higher AI literacy, while those who applied suggestions uncritically demonstrated lower AI literacy regardless of the amount of AI feedback they received. This study offers practical recommendations for students and lecturers on the wise and critical use of AI-generated feedback in the proposal and thesis writing process.

The COM Essay Assessor

The development and calibration of the COM Essay Assessor is presented, a rubric-based generative artificial intelligence (GenAI) tool designed to support formative feedback while retaining instructor oversight and reflects on the opportunities and challenges of integrating GenAI into large writing programs.

Teaching writing when the sentences sound better than the thinking

The paper shows that AI has not displaced teacher feedback but clarified its interpretive, relational and ethical value in English teaching, and points to the need for further classroom-based work on feedback, AI-assisted revision and multilingual writing.

Doing DDL with AI: Classroom Interactions in Collaborative Writing

Abstract In the last few decades, data-driven learning and generative artificial intelligence (AI) applications have become increasingly visible in academic writing instruction. The data-driven learning approach promotes students’ lexical and morphological awareness by encouraging them to explore language in context, while generative AI tools offer new possibilities for idea generation, text development, and linguistic feedback. However, limited process-based studies exist on how these two approaches work together in collaborative writing in the classroom and how teacher guidance shapes this process. This study aims to explore how B2-level students enrolled in a university preparatory program use data-driven learning and generative AI tools in the collaborative opinion essay writing process. Designed as a qualitative classroom-based case study, the research was conducted over a seven-week module. Structured student diaries, teacher journals and semi-structured interviews were used in the data collection process. Findings indicate that tool use was constructed through peer negotiation and teacher mediation. Students used generative AI for idea generation and corpus tools for lexical verification and contextual checking. The study highlights the critical role of pedagogical design and teacher facilitation in DDL-AI-assisted writing instruction.

Student Evaluation of Repeated AI Feedback Across a Semester of Writing

Analysis of reflective essay-feedback-appraisal instances from 283 Estonian bachelor students across one semester contributes descriptive classroom evidence on integration of AI feedback - a fast and scalable way to provide immediate writing advice, but not a self-contained route to better reflection.