π€ Responsible AI Usage
ArcLycΓ©e demonstrates how AI can be used as a pedagogical tool under constant educator supervision, with students always in control of decisions.
Human-Led Collaborative Design
The game was designed entirely by the students of les fous du robot (~13 years old) at the Liceo FrancΓ©s de Santo Domingo, under the supervision of their robotics teacher, assistants and external advisors. The process included:
- Group discussion sessions β Students explored, debated and decided on every aspect of the game: worlds, characters, mechanics, narrative and tone.
- Research assignments β Each student researched a specific topic (archaeology, fauna, laws, historical figures) and presented findings to the group.
- Real visits and interviews β Visits to heritage sites and interviews with experts were completed as part of the preparation.
- Design document β A comprehensive document was edited and revised 10 times between February 10 and March 13, 2026, before development began. View original document.
Students in the Game
The game includes every team member as a tangible presence:
- Interactive NPCs β All 10 students appear as characters in the LFSD world (a replica of their robotics classroom), with dialogues reflecting their real personalities and interests.
- Their teacher β Prof. Nicolas Droulers is present as a mentor with rotating dialogue about robotics and heritage.
- Personalities reflected β Each student's gaming preferences, skills and personality are manifested in their dialogues, the mini-games they propose and the easter eggs they hide.
- The real lab β The LFSD world faithfully replicates the robotics classroom: tables with Scratch, FIRST LEGO League mat with animated robot, 3D printer.
AI as a Development Tool
Once the design document was complete, AI (Claude by Anthropic) was used as a tool to:
- Implement documented code β Vanilla JavaScript with extensive comments in Spanish, readable by 13-year-old students.
- Generate procedural art β Sprites and scenarios generated programmatically (no external artist dependencies).
- Create original music β Musical prompts designed by students, generated by AI (Suno) with metadata crediting the team.
- Develop documentation β Trilingual informational website with technical, pedagogical and content documentation.
- Organize and track changes β Git as version control, detailed CHANGELOG, basic CI/CD.
Human Supervision and Control
At all times, AI usage was supervised and controlled by humans:
| Aspect | Human Role | AI Role |
| Narrative and content | Students decide themes, characters, tone | Implements decisions in code and text |
| Game mechanics | Students propose and test | Programs logic per specifications |
| Historical accuracy | Students research, advisors verify | Integrates verified data into the game |
| Balance and fun | Students test and adjust | Implements requested adjustments |
| Code quality | Teacher reviews architecture and patterns | Writes readable code with comments |
| Publication | Teacher and students approve each release | Generates changelogs and updates documentation |
Testing and Improvement Cycle
Once the first playable prototype was available:
- Internal testing β Students played the game exhaustively, reporting bugs and suggesting improvements.
- External testing β Shared with friends, family and other Liceo students for fresh perspectives.
- Group discussion β Suggestions were discussed as a team, prioritizing those that improved both the educational and play experience.
- Iterative implementation β Each approved improvement was implemented, tested and deployed, with students verifying the result.
- Continuous improvement β The cycle repeats: new ideas emerge from testing, are discussed, implemented and verified.
Ethical Principles of AI Usage
The project establishes and demonstrates these principles:
- Transparency β AI usage is publicly documented. Every commit states "Co-Authored-By: Claude". There is no attempt to hide the tool.
- Human agency β Students make all creative and content decisions. AI executes, it does not decide.
- Teacher oversight β The teacher and advisors review AI usage, ensure content is appropriate and that students understand what is generated.
- Understanding over delegation β Students can explain how every part of the game works. It is not a black box.
- Attribution β The student team is credited as creators. AI is credited as a tool, just as one would credit a compiler or an IDE.
Skills Developed
Using AI as a tool (not a replacement) has allowed students to develop:
- Critical thinking β Evaluating and verifying AI outputs, rejecting what is incorrect, requesting adjustments.
- Precise communication β Learning to formulate clear and specific instructions (prompting as a communication skill).
- Project management β Coordinating a complex project with multiple phases, deliverables and stakeholders.
- Digital literacy β Understanding what AI can and cannot do, its limitations and biases.
- Technology ethics β Reflecting on responsible use of AI tools in creative and educational contexts.
Summary
ArcLycΓ©e is an example of responsible and pedagogical AI usage: students lead, decide, test and approve. AI accelerates the implementation of their ideas without replacing their creativity, judgment or learning. The result is a game that is genuinely theirs β in content, in vision and in spirit β empowered by modern tools they learned to use with discernment and under proper supervision.