Theory-Guided Human-AI Collaboration: Focusing, Feedback and Critical Examination in Jean-Pol Martin's Semantic Field
This methodological article examines how Jean-Pol Martin organizes theory-guided collaboration with artificial intelligence through two complementary documents: a theory prompt and a master prompt. The theory prompt supplies the conceptual framework of Learning by Teaching (LdL), the New Human Rights (NMR), and Martin’s semantic field. The master prompt specifies analytical procedures, working expectations, and criteria for provisional claims. Ongoing human feedback connects these instructions to the concrete task. The article’s contribution lies in making the relationship between conceptual grounding, procedural direction, and situated correction explicit. Two episodes from the dialogue leading to this publication illustrate how editorial feedback and the introduction of foundational documents reshape an AI-generated text. Within the developing series on LdL, NMR, and human-AI collaboration, the article functions as a methodological bridge between theoretical exposition and practical application. It offers a reusable procedure for documenting instructions, corrections, and revisions, together with suggestions for comparative evaluation. The organismic reference point of the NMR framework places the living, thinking, social, and meaning-oriented human being at the centre of analysis. Six equal-ranking needs provide its anthropological and normative orientation: thinking, health, security, social belonging, self-realization and participation, and meaning. This is a methodological account with illustrative dialogue episodes, not a controlled study. It distinguishes conceptual coherence from factual validation and practical effectiveness. The article was developed through collaboration between Jean-Pol Martin and ChatGPT. Supplementary materials: The deposit includes the original German master prompt and theory prompt with its organismic supplement. The main article is in English; the supplementary documents are in German.
Authors
- Jean-Pol Martin (ORCID: https://orcid.org/0000-0002-9998-2432)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-02
- DOI
- https://doi.org/10.5281/zenodo.23886393
- Primary Topic
- Qualitative Research Methods and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00