Keynote speakers

JIN Tan, PhD

Professor

School of International Culture

South China Normal University

Keynote Title

Scaffolding Teacher Agency in Collaborating with GenAI for Text Simplification: A Three-Phase Support Framework for Materials Development

Bio

Professor Tan Jin is Dean of the School of International Culture, South China Normal University, China. His research interests include corpus linguistics, language assessment, and computer-assisted language learning. He leads a research team that has developed LanguageData.net, an integrated intelligent platform for language teaching and research. The platform’s bilingual Chinese–English text complexity analyzer has attracted registered users from more than 15,000 institutions worldwide. He serves as Associate Editor of Language Testing in Asia and sits on the editorial boards of journals including Innovation in Language Learning and Teaching and Journal of English for Academic Purposes.

Abstract

Text simplification is central to language materials development, aligning authentic texts with learners’ proficiency levels. While the recent advent of generative artificial intelligence (GenAI) has enhanced the efficiency of text simplification, teachers often rely on intuition-led prompting and accept outputs that do not fully match curricular goals or learner needs, which underscores a pressing need for scaffolding teacher agency, that is, the capacity to make informed, intentional decisions in AI-mediated contexts.

This keynote presents a three-phase support framework designed to systematically scaffold teacher agency in GenAI-assisted text simplification for materials development. Phase 1, data-driven tool development, centers on a corpus-informed analyzer that transforms lexical and syntactic complexity into transparent, data-driven diagnostics. This evidence-based diagnosis empowers teachers to pinpoint simplification needs and ground their prompting in evidence, which supports personal agency by enabling teachers to explain and justify their decisions. Phase 2, peer-mediated training design, engages teachers in structured workshops centered on peer inquiry. Through joint prompt engineering, reciprocal critique, and co-design activities, teachers collectively evaluate GenAI outputs and align them with curricular goals, which cultivates relational agency and fosters a professional community where pedagogical expertise is negotiated and refined through dialogue. Phase 3, school-based network building, extends professional learning through experimental school networks. Trained teachers create, pilot, and disseminate context-specific micro-lessons and adapted materials across a digital platform, which fosters collective agency by supporting the co-construction of professional standards for responsible GenAI integration for text simplification beyond individual classrooms. Preliminary implementation indicates that this integrated framework improves prompt precision, alignment with learner needs, and teacher engagement.

Overall, the three-phase framework has provided systematic pedagogical scaffolding for teacher agency, advancing theoretically informed, teacher-centered human–AI collaboration to optimize text simplification and elevate the quality of language materials development. Theoretically, the framework has advanced understanding of teacher agency in human–AI collaboration by conceptualizing agency as a multi-layered and scaffolded condition that develops across individual, interpersonal, and institutional levels. Practically, the framework has offered a replicable model for language teacher education by linking tool development, teacher training, and school-based networks in ways that keep pedagogical judgment central to GenAI-assisted materials development. These findings have implications for designing professional support that helps teachers use GenAI responsibly and effectively across diverse instructional contexts.

Disclaimer

Site Map