A Cybernetic Perspective on GenAI-Enhanced Feedback in Higher Education: A Thing Interviewing Exploration Protocol

Description

This repository contains the supplementary materials supporting the study A Cybernetic Perspective on GenAI-Enhanced Feedback in Higher Education: A Thing Interviewing Exploration. The dataset comprises four semi-structured thing interview transcripts conducted with widely used generative artificial intelligence (GenAI) systems, a coding framework and a theme development matrix, and representative excerpts supporting the thematic analysis. The interviews explored how GenAI systems construct and represent their participation in educational feedback processes when examined through the lenses of first-order cybernetics, second-order cybernetics, and radical constructivism. The materials document the analytical process from data collection to theme development and are provided to enhance transparency, traceability, and methodological rigour. Due to the evolving nature of GenAI systems, exact replication of interview responses may not be possible; however, the repository provides the complete dataset and analytical documentation used in the reported study.

Steps to reproduce

1. Develop an interview protocol informed by educational feedback theory, first-order cybernetics, second-order cybernetics, and radical constructivism. The protocol used in this study is provided in Appendix A of the article. 2. Conduct independent semi-structured thing interviews with multiple GenAI systems using the same interview protocol. In this study, interviews were conducted with ChatGPT, Gemini, DeepSeek, and Claude. All systems were instructed to respond in the first person and describe their participation in educational feedback processes. 3. Record and save the complete interview transcripts. The transcripts used in this study are provided as Supplementary Materials S1–S4. 4. Analyse the transcripts using thematic analysis (Braun & Clarke, 2006). Initial coding should be guided by the analytical constructs derived from the cybernetic framework: observation, comparison, regulation, adaptation, structural coupling, and reflexive observation. 5. Review coded segments iteratively to identify recurring patterns, relationships, and representations across the dataset. Develop themes through comparison across transcripts and researcher discussion. 6. Refine and interpret the themes using the coding framework and theme development matrix provided in Supplementary Material S5. 7. Compare the resulting themes with the representative excerpts provided in Supplementary Material S6 and interpret the findings through the theoretical framework presented in the article. Note: Because GenAI systems are continuously updated and may generate different responses over time, exact replication of interview outputs cannot be guaranteed. The materials provided in this repository document the dataset and analytical process used in the reported study (May-June 2026).

Institutions

• Universidad Panamericana

  Mexico City, Mexico City
• Aston University

  England, Birmingham
• Tecnológico de Monterrey

  Nuevo León, Monterrey

Categories

Licence