Prometheus: From Generic Chatbots to Frugal, Ethical and Socratic AI for Education

Team name: Prometheus

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

• Which tool(s) did you use? Gemini for images, Prometheus AI for feedback (Socratic mode)

• For what purpose(s)? (e.g., research, structuring ideas, drafting, data analysis, etc.) Image generation and feedback

• How did these tools support your work? We wrote the first draft without AI with our own reflection and feedback from comments and keynotes then used Prometheus AI (Socratic mode) to improve our content (for example, it made us think about the fact we could use the allegory of a Constitution and laws to describe our expansion)

External feedback & contributions : Please list the people who supported you during the process (e.g., experts, professors, professionals):

• Survey respondents - Prometheus users - Feedback - Allowed us to understand how Prometheus AI is used

• Ndeye Sophie Seck, Josela Aramburu - Commentator - Advised us on usage tests

• Lucie Jacquet Malo - Commentator - Advised us on national expansion and funding

• Caroline Beslin - Associate Lecturer in Project Management and Communication at Centrale Lyon - For her keynote on how to create a successful project that inspired us and helped us identify areas for improvement in our project.

• Alexandra Prégent - PhD Researcher at CEA - The keynote on Ethics and AI sparked some debate and helped us re-assess our project against the European ALTAI framework and new ethical considerations.

• Saiph Savage - Assistant Professor, Director, Civic A.I. Lab at Khoury College of Computer Sciences - Insights on the keynote on how to build AI for people helped us improve our project's inclusivity. 

Honorable mentions:

These comments did not inspire new ideas but contributed to the reflection on the project and we would like to thank the respondents:

• Youri from the ÒCTAVIA Team

• Johanna F from the T-Twice Team

• Nipun Ranchhod Navadia

• Hans Zúñiga

Initial contribution: Prometheus: From Generic Chatbots to Frugal, Ethical and Socratic AI for Education

Final contribution: 

Our final contribution can be accessed here:

Reflection on the process

This reflection is detailed in the final contribution video.

Two sources fundamentally reshaped Prometheus during Phase 2:

• Expert keynotes provided frameworks we didn't know existed. Alexandra Prégent introduced the ALTAI framework (Europe's seven-pillar standard for trustworthy AI). We discovered our approach already complied with formal ethical requirements. Dr. Savage's bias research pushed us to diversify cultural references and plan multilingual expansion. Alain Goudey's Volition Advantage research also reinforced our choice of developing our "Socratic mode".

• Community feedback: user satisfaction surveys revealed unexpected patterns. Our 150 students' API usage spikes before exams, not during assignments: they are thus using Prometheus to revise and understand, not to generate work. This validated our approach but showed we needed better onboarding about learning with AI versus outsourcing to AI. We reframed our positioning around objective standards rather than vague ethics claims: thus, instead of "we care about ethics," we can now proudly say that we comply with existing frameworks and plan to draft our own "Constitution" with your contributions.

Thus, Phase 2 transformed intuition into infrastructure: this feedback process first provided us with a solid set of references to anchor our ethical, sustainable, and frugal approach with recognized standards rather than intuitive principles. Second, it equipped us with the vocabulary and credibility necessary to present Prometheus to institutional stakeholders (universities, regulatory bodies, potential partners) at national and international scales. Instead of positioning ourselves as "students with good intentions," we can now demonstrate compliance with European AI ethics standards, cite peer-reviewed research, and reference established pedagogical methodologies. This transformation from informal project to theoretically grounded initiative makes scaling feasible and defensible.