Teacher authority, responsibility, and platform governance in generative AI-supported classrooms

Abstract Public debate on large language models (LLMs) in education often asks whether artificial intelligence will replace teachers. This article argues that this question misses a more important shift: in postdigital classrooms, teachers’ authority is not disappearing but being reorganized. As generative AI reduces the centrality of teachers’ authority based on knowledge possession, teachers are increasingly expected to exercise three other forms of authority: epistemic-arbitration authority, pedagogical-orchestration authority, and normative-accountability authority. They must judge the reliability of AI outputs, decide when and how AI should be used, and uphold standards of fairness, academic integrity, public reasoning, and student safety. Drawing on recent scholarship on teacher agency, professional judgement, AI literacy, assessment, responsible AI, and platform governance, the article argues that this shift is accompanied by a downward transfer of responsibility. While platform companies and model providers gain infrastructural power over information flows, curricular resources, and interface conditions, teachers remain the proximate actors held responsible for verification, inclusion, safeguarding, and educational purpose. To capture this condition, the article introduces judgement labour as a labour-process concept that names how professional judgement is repeatedly required, intensified, unevenly distributed, and often left unrecognized through the work of verification, orchestration, translation, refusal, ethical mediation, and continuing AI learning. The article concludes that the key issue is not teacher obsolescence, but an authority-responsibility asymmetry.