AI as a Critical Opponent: Effects on L2 Writers’ Warranting and Argument Quality
Abstract:
Assessing L2 students’ argumentative writing quality requires attention not only to the presence of key components including thesis, evidence, counterargument, and rebuttal, but also to the depth of reasoning that connects these elements. Many EFL students can produce structural components yet struggle to justify their claims clearly. This study investigates whether AI-mediated dialogic practice via ChatGPT 5.1 can improve argument quality in a B2-level writing course. In a quasi-experimental design, both groups receive instruction on argumentation, but the experimental group also engages in AI-supported pro–opponent exchanges in which the AI challenges claims, questions evidence, and presents counterarguments for students to rebut. Essays from both groups will be rated with an analytic rubric aligned with CEFR principles and adapted through stakeholder input. By foregrounding assessment outcomes, the study examines the potential of dialogic AI to support instructional design and evidence-based evaluation in argumentative writing.
Date:
16.02.2026, 16:00 UK time
Speakers:
Burcu Kayarkaya
Selahattin Yılmaz
Burcu Kayarkaya, PhD, is an EAP instructor in the Modern Languages Department at Yıldız Technical University. She also works as a language test developer and item writer/editor for a high-stakes exam body and serves as a consultant for the EU–Council of Europe joint project on enhancing foreign language education quality in Türkiye. She holds an MA in ELT, and her PhD focused on translanguaging-informed writing instruction and assessment design. She has presented in various international conferences, including IATEFL 2025, where she was awarded the TOEFL scholarship.
Selahattin Yılmaz, Ph.D. is an EAP instructor at Yıldız Technical University in Istanbul and currently serves as a language test developer for the EU–Council of Europe Joint Project on foreign language education quality in Türkiye. He completed his Ph.D. in Applied Linguistics at Georgia State University and has published in leading journals such as English for Specific Purposes, Journal of English for Academic Purposes and Porta Linguarum. His research focuses on second language writing, corpus-based analysis, and AI-mediated pedagogy, and he has presented at major international venues including AAAL, AACL, and EALTA
TEASIG – Automated Assessment of L2 Learners’ Written Production: Can AI Make a Difference?
January 7
The development of writing skills constitutes a core component of second language education, yet the assessment of learners’ written production remains a persistent challenge. Providing reliable scores and informative feedback is both time-intensive for teachers and problematic in terms of consistency in high-stakes examinations. Over the past decades, automated assessment systems have sought to address these issues. They were driven by advances in natural language processing and artificial intelligence. However, until recently such systems could only be developed by highly specialised teams for narrowly defined tasks, limiting their adaptability and accessibility for classroom use. The recent emergence of Generative AI and chatbots marks a turning point, offering accessible tools for automated assessment and feedback. This talk reviews the evolution of automated systems for L2 writing evaluation, examines the opportunities afforded by GenAI-based chatbots, and presents recent empirical findings on their effectiveness and pedagogical implications.
Speaker:
Agnieszka LEŃKO-SZYMAŃSKA
Agnieszka LEŃKO-SZYMAŃSKA is an Associate Professor at the Institute of Applied Linguistics, University of Warsaw. Her research interests lie at the intersection of second language acquisition, language teaching and assessment, and corpus linguistics, with a particular focus on lexis and phraseology. She is actively involved in teacher education and professional development. Her monograph Defining and Assessing Lexical Proficiency (Routledge, 2020) explores the construct of word use in L2 writing. She has also published widely on the applications of corpora in language teaching, learning, and assessment.