The Prompt is Not Enough: Rethinking AI-Learner Interaction in the Post-Prompting Era

Abstract

The rise of generative AI marks a fundamental shift in human-AI interaction, where learners are no longer limited to prompting static tools but engage dynamically with agentic AI partners. This conceptual paper introduces the notion of the Post-Prompting Era, characterised by AI systems that co-regulate learning, scaffold reflection, and evolve across dialogues. Moving beyond prompt engineering, we argue that both teachers and students must develop new forms of interactional and co-regulatory literacy. We examine how this shift redefines pedagogical roles, learning task design, and human-AI orchestration. Practical implications include rethinking teacher professional development, lesson planning, and techno-pedagogical co-design with AI agents. We also outline a research agenda that calls for interdisciplinary inquiry into the cognitive mechanisms, methodological innovations, and ethical boundaries of agentic AI in education. While emerging features such as ChatGPT's Study Mode offer promising instantiations, our analysis cautions against oversimplified automation. Instead, we advocate for a nuanced reconceptualisation of AI's role as a co-participant in learning – one that complements human agency rather than replacing it.