Teaching Materials & Implementation Guide
The A.U.D.I.T. Protocol consists of four core sequences designed for secondary education (ages 12-18). Each sequence is available in versions adapted for both middle school and high school students, with differentiated complexity and cognitive demands.
Complete detailed resources (student worksheets, teacher guides, assessment rubrics, and downloadable PDFs) are currently available in French. This page provides an overview of the pedagogical approach and structure.
Students develop a practical ability to recognize AI-generated content, understanding that AI has distinctive stylistic "fingerprints" even when mimicking human writing.
Guide students toward discovery rather than lecturing. The power comes from students identifying patterns themselves.
Students discover firsthand that AI can "hallucinate" facts with complete confidence. This visceral experience is far more powerful than being told "AI isn't always accurate."
Safety in intellectual risk-taking. Students should feel free to experiment without judgment. The goal is discovery, not getting "right answers."
Students grasp the fundamental limitation—AI predicts next words, it doesn't "think." This understanding empowers better tool use.
The charter creation is crucial. Let students develop their own rules based on discoveries. Autonomy is the goal.
Students recognize that AI reflects and can amplify societal biases. This creates critical consumers of AI, not passive acceptors.
Handle sensitively. This sequence touches on representation, stereotypes, and fairness. Create a safe space for honest discussion.
Students understand that AI security is an ongoing challenge, not a solved problem. They recognize manipulation techniques not to use them maliciously, but to think critically about AI safety claims.
Ethical framing is crucial. This sequence should be presented as security awareness education, similar to teaching about phishing emails. The goal is defensive knowledge, not offensive capability.
All sequences share a common methodology:
Students discover AI limitations through direct experience, not lectures. The teacher orchestrates experiments, not explanations.
Students move from concrete cases ("Look, this AI output is wrong!") to general principles ("AI can hallucinate facts").
Rather than imposing rules, students create their own based on discoveries. Ownership creates adherence.
Sequences are designed to be adapted. Teachers should modify based on student responses and local context.
Duration: 1 class period (45-90 minutes)
Impact: Introduction to AI critical thinking
Recommended: Sequence 02 (trap-setting) for immediate engagement
Duration: 4 class periods over 2-4 weeks
Impact: Complete A.U.D.I.T. framework
Recommended: Sequential order (01 → 02 → 03 → 04)
Duration: Initial 4 periods + ongoing integration
Impact: Embedded critical AI literacy across curriculum
Recommended: Return to A.U.D.I.T. framework whenever using AI tools
While developed for mathematics, A.U.D.I.T. applies across subjects:
The protocol is designed to be culturally transferable:
For teachers comfortable in French or with translation tools, the complete resource package includes:
If you're implementing A.U.D.I.T. in an international context and need guidance:
These sequences evolve based on classroom experience. If you test them:
Your observations help refine the protocol for international use.
Timeline: Based on user interest and feedback
The A.U.D.I.T. Protocol isn't about memorizing rules—it's about developing a critical mindset. Students who complete these sequences don't just learn "AI can be wrong." They learn to:
This is the "Columbo" stance—critical validator, not passive receiver.
Ready to implement?
Start with Sequence 02 (trap-setting) for immediate student engagement, or proceed sequentially through all four for complete critical AI literacy development.
Questions? Contact: philippe.dupeyrat@icloud.com
Full materials (French): philipped79.github.io/audit-ia