A Good Prompt Is More Than an Instruction: Jean-Pol Martin's Master Prompt as a Framework for Human-AI Collaboration

This working paper examines Jean-Pol Martin’s Master Prompt as a theoretical and practical framework for human-AI collaboration. It argues that a good prompt is not merely a technical instruction, but a semantic and normative infrastructure that guides AI analysis. Based on Martin’s semantic field — Learning by Teaching, New Human Rights, basic needs, thinking, coherence, flow, participation, meaning, and life preservation — the paper shows how prompt architecture can structure AI interaction in ways that support human agency, systemic analysis, and constructive action.