Co-creation students and AI, Cognitive Trajectory and Epistemic Governance of AI in Higher Education

Team name: Eidofüge

Use of AI tools : If you used AI, please specify:

• Which tool(s) did you use? Claude (Anthropic), Gemini (Google), DeepL and other, that are directly integrated into tools.

• For what purpose(s)?

  • Research and state of the art : surveying the academic literature on cognitive offloading, AI in education, and the AIOLIA framework, then summarising long articles to extract the points relevant to our argument.

  • Translation : moving fluidly between French (our working language) and English (the submission language) without losing nuance, via DeepL.

  • Structuring ideas : testing the logical coherence of our framework (See / Learn / Decide), checking whether each recommendation R0–R9 was internally consistent, and stress-testing our claims.

  • Critical feedback : using the models as a complement to the feedback received from other participants and mentors, to surface counter arguments we may have missed.

• How did these tools support your work?

  The tools accelerated tasks where speed mattered (literature scanning, translation, formatting) and provided an additional layer of critical dialogue alongside human feedback. They did not replace the team's thinking: the four student cases, the choice of cognitive trajectory as the central concept, the architecture of the framework, the recommendations and their political stance were debated and decided by the team. The AI was a working partner for execution and verification, fully consistent with the very thesis of our proposal, which argues not against AI use, but for making such co-creation visible, deliberate, and accountable.

External feedback & contributions :

• Name : Bruno De Lièvre

  Role : Professor, Faculty of Psychology and Educational Sciences, University of Mons (Belgium)

  Type of contribution: Detailed written feedback on the initial proposal (structure, methodological rigour, safeguards, pilot design).

• Name: Nipun Ranchhod Navadia

  Role: Participant, AI Grand Challenge

  Type of contribution: Critical feedback on the operational gap of the initial proposal, prompting the formalisation of the autonomy gap (Δauto) and the three complementary indicators.

• Name: Chinmay Das

  Role: Participant, AI Grand Challenge

  Type of contribution: Early feedback on the need for practical measurement of critical thinking and autonomy in AI-supported environments

• Name: Caroline Beslin

  Role: Associate Lecturer in Project Management and Communication, Centrale Lyon

  Type of contribution: Keynotes Project Success 1 & 2 (23/04 and 04/05) four-dimension SWOT framing and the five stages of knowledge transmission, which informed our five-stage Cognitive Mirror Protocol

• Name: Alexandra Pregent

• Role: PhD Researcher, CEA Societal Impact of AI, Information Ethics, Privacy

• Type of contribution: Keynote Ethics of AI (28/04) introduction to the AIOLIA project methods and ethical dilemmas, which led us to anchor Eidofüge to the AIOLIA framework

• Name: Alain Goudey

  Role: Associate Dean for Digital, NEOMA Business School

  Type of contribution: Keynote What 81,000 people (didn't quite) tell us about AI (05/05) empirical perspective on real-world AI use, which shaped our reading of usage modes (healthy / instrumental / delegation)

• Name: Saiph Savage

  Role: Assistant Professor, Director of the Civic AI Lab, Khoury College, Northeastern University

• Type of contribution: Keynote Human-Centered AI (06/05) design principles that confirmed our Human-Centered AI framing

• Name: Bastien Guerry

  Role: Independent programmer, open-source contributor, digital commons

  Type of contribution: Keynote AI Coding Assistants and Free Software (07/05) informed our choice of a frugal, sovereign, fully open-source technical stack and our attention to energy footprint.

• Name: Chelza

  Role: AI Grand Challenge community moderator

  Type of contribution: Contextual support during exchanges with experts on the platform

• Name: Team Eidofüge (EPHEC)

• Role: Four students from Health Technologies, e-Business, International Business and Law

  Type of contribution: Internal deliberation, shared writing, peer review of every section, lived cases that grounded the framework .

  Beyond these named contributors, the broader exchanges within the AI Grand Challenge community (peer comments, discussion threads, cross-team observations) provided a continuous environment of critique and refinement, for which we are sincerely grateful.

Initial contribution: Co-creation students and AI, Cognitive Trajectory and Epistemic Governance of AI in Higher Education

Final contribution: https://drive.google.com/file/d/1xeMKI3rnQnONDqiTPytdLD4oNTXRqPj-/view

Reflection on the process (important)

Explain how your contribution evolved during Phase 2:

• Feedback that influenced our work.Written feedback from Prof. Bruno De Lièvre (UMons) on the initial proposal , asking for a clearer case then problem follow by framework finally implementation structure, more cautious claims, an explicit method with indicators, named safeguards, and a pilot. Critical comments from participants Nipun Navadia ("the proposal remains diagnostic, an operational instrument is missing") and Chinmay Das (practical ways to measure autonomy). Keynotes by Alexandra Pregent (AIOLIA), Caroline Beslin (project success and the five stages of knowledge transmission), Alain Goudey (real-world modes of AI use), Saiph Savage (Human-Centered AI), and Bastien Guerry (open-source, sovereignty, sustainability).

• The final version follows the requested structure; "generational risk" is replaced by cognitive trajectory with four longitudinal indicators (autonomy gap Δauto, justification, error detection, transfer); a five-stage Cognitive Mirror Protocol is introduced; the framework is anchored to AIOLIA via a matrix mapping each ethical principle to a concrete safeguard; R7 and R9 are reframed so the Global-South dimension is led by the communities concerned; a three-phase deployment roadmap with pilot is added, along with a frugal, fully open-source technical stack and an energy/CO₂ accounting layer.

• How did this process strengthen your final proposal?

  How the process strengthened the proposal.External feedback turned a diagnosis into a deployable instrument: claims are now evidence-based, the method is testable, ethics are auditable against a published reference, and implementation is concrete and staged. The proposal is shorter in ambition-per-sentence and stronger in operational substance.

Video pitch

Please provide a link to your 1-minute video pitch presenting your work: https://drive.google.com/file/d/1PsnH8OHvNXu8fx9eWEoHCLr5iJE9GVdU/view