The JUDGE Model of AI-Mediated Teacher Agency: A Theoretical Framework for English Language Teaching

Although generative artificial intelligence (AI) has rapidly penetrated English language teaching (ELT), the existing research literature has remained largely limited to technocentric views of usability and has ignored a critical theoretical gap in the reconfiguration of teacher agency and professional authority. Aiming to address this gap, this article introduces the JUDGE model, first conceptualized by Hosseini (2026), as a leading framework for analyzing teacher agency in AI-mediated ecosystems. Drawing on sociocultural theory and ecological perspectives, Hosseini’s JUDGE model identifies five critical dimensions of professional action: Justification, Uncertainty Management, Decision Ownership, Guidance vs. Governance, and Ethical Accountability. This paper moves beyond abstract theorizing to explicitly operationalize this framework by introducing the JUDGE Scale (or J-Scale). Preliminary validation of the J-Scale by experts confirms its high content validity and transforms the dimensions of the JUDGE model into measurable constructs for empirical inquiry. By simultaneously providing a conceptual architecture and a valid psychometric trajectory, this study provides a solid and practical foundation for researchers and teacher educators to navigate the redistribution of agency in the age of artificial intelligence.Keywords: Artificial intelligence in ELT; Teacher agency; AI-mediated instruction; Professional judgment; English language teaching