Using Chatgpt “Teacher Mode” To Teach: A Design-Based Research Protocol and Implementation Framework

Abstract

The unstructured use of generative artificial intelligence (AI) in education may encourage passive answer-seeking rather than active learning. This study introduces ChatGPT Teacher Mode, a structured framework designed to position generative AI as a guided instructional partner that promotes critical thinking, metacognition, and evidence-based reasoning. Using a design-based research approach, the study develops a protocol for implementing Teacher Mode across mathematics, history, civics, and geography. The framework incorporates discipline-specific learning activities, including articulating mathematical reasoning, evaluating primary sources, explaining constitutional principles, and interpreting climate maps. It integrates pedagogical scaffolding, formative assessment, verification routines, and transparent AI-use policies to support responsible implementation. The proposed framework targets four outcomes: improved learning through structured reasoning, enhanced metacognitive awareness, higher-quality feedback aligned with assessment rubrics, and greater transparency in AI use. Potential challenges, including inaccurate outputs, over-reliance, and inequitable access, are addressed through verification strategies, AI-assisted and AI-free assessments, and institutional guidelines. This study contributes a socio-technical framework, an empirically testable research protocol, and practical guidance for educators integrating generative AI into teaching and learning. Teacher Mode offers a structured approach to leveraging generative AI while maintaining academic integrity and supporting meaningful student engagement. Its effectiveness, however, requires empirical validation across educational contexts.