Allocating epistemic responsibility in generative AI-supported scientific inquiry: a teacher-orchestrated multi-agent framework

Journal: Region - Educational Research and Reviews DOI: 10.32629/rerr.v8i5.5479

Ziyun ZHANG

Tianjin University of Technology and Education

Abstract

Generative AI can supply an explanation before a learner has decided what evidence would make it credible. In scientific inquiry, AIenabled assistance may improve an answer while removing responsibility for judging it. This paper proposes epistemic responsibility allocation as a guiding principle for educational multi-agent systems. A responsibility specifies what an actor must justify, what the actor is prohibited from doing, and when decisionmaking authority passes to a teacher. An Expert verifies approved sources, a Peer raises evidence-seeking challenges, and a Coach manages and withdraws instructional scaffolds. A ledger adapts Toulmin's structure to record claims, evidence, warrants, qualifiers, challenges, and student revisions. Four propositions are examined through a classroom comparison with a single-agent interface, providing an initial empirical test of the framework. Multiple agents are useful only when permissions create distinct, auditable responsibilities among them.

Keywords

generative artificial intelligence; epistemic responsibility; scientific argumentation; multi-agent systems; learning design; teacher orchestration

References

[1] Chiu T K F. A classification tool to foster self-regulated learning with generative artificial intelligence by applying self-determination theory: A case of ChatGPT[J]. Educational Technology Research and Development, 2024, 72: 2401-2416. https://doi.org/10.1007/s11423-024-10366-w
[2] Lee H Y, Chen P H, Wang W S, et al. Empowering ChatGPT with guidance mechanism in blended learning: Effect of self-regulated learning, higher-order thinking skills, and knowledge construction[J]. International Journal of Educational Technology in Higher Education, 2024, 21: 16. https://doi.org/10.1186/s41239-024-00447-4
[3] Hmelo-Silver C E, Duncan R G, Chinn C A. Scaffolding and achievement in problem-based and inquiry learning[J]. Educational Psychologist, 2007, 42(2): 99-107. https://doi.org/10.1080/00461520701263368
[4] Alfredo R, Echeverria V, Jin Y, et al. Human-centred learning analytics and AI in education: A systematic literature review[J]. Computers and Education: Artificial Intelligence, 2024, 6: 100215. https://doi.org/10.1016/j.caeai.2024.100215
[5] Miao F, Cukurova M. AI competency framework for teachers[R]. UNESCO, 2024.
https://www.unesco.org/en/articles/ai-competency-framework-teachers
[6] Chi M T H, Wylie R. The ICAP framework: Linking cognitive engagement to active learning outcomes[J]. Educational Psychologist, 2014, 49(4): 219-243. https://doi.org/10.1080/00461520.2014.965823
[7] Toulmin S E. The uses of argument (Updated ed.)[M]. Cambridge: Cambridge University Press, 2003.
[8] Osborne J, Erduran S, Simon S. Enhancing the quality of argumentation in school science[J]. Journal of Research in Science Teaching, 2004, 41(10): 994-1020. https://doi.org/10.1002/tea.20035
[9] The Design-Based Research Collective. Design-based research: An emerging paradigm for educational inquiry[J]. Educational Researcher, 2003, 32(1): 5-8. https://doi.org/10.3102/0013189X032001005

Copyright © 2026 Ziyun ZHANG

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License