GenAI-Assisted Language Teaching (GenAILT): Approach, Design, and Procedure

Since the release of ChatGPT in late 2022, Generative Artificial Intelligence (GenAI) has been rapidly taken up in second language teaching and learning. Much of the existing literature reports on perceptions, tools, techniques, and short-term outcomes; however, the broader question of what kind of pedagogy is emerging has received less attention. This conceptual paper addresses that question by analysing GenAI-assisted Language Teaching (GenAILT) through Richards and Rodgers’ (2014) framework of Approach, Design, and Procedure. At the level of Approach, we demonstrate that GenAI is deployed under quite different theoretical assumptions, and we distinguish declared theory from the commitments GenAI enacts through its training data, which favour dominant language varieties unless teachers counter them. At the level of Design, we argue that objectives in GenAILT are better understood along a product–process continuum, that GenAI functions as a flexible resource for realising existing syllabus types, and that activity design must build in strategic, critical and ethical engagement with GenAI output. We further propose a four-role matrix for learners along the dimensions of learning orientation and instructional scaffolding, examine how teacher work is redistributed, reframe instructional materials in terms of form and quality assurance, and discuss classroom assessment as a design decision. At the level of Procedure, we illustrate that competent GenAILT depends on teacher decision-making. We close with implications for teacher education and research.