Aletheia Institute

Why Aletheia?

Everyone now agrees on one obvious thing: students and teachers must be trained for artificial intelligence.

But training for AI using the slow, rigid, top-down methods of a pre-AI world would be a contradiction. AI isn't just a technological revolution: it's a cognitive shift. It transforms the very way we search, learn, understand, produce and assess knowledge.

Education likes to take its time. Artificial intelligence, by its very nature, carries acceleration. You won't train people for a paradigm shift using the habits of the previous paradigm.

That's why Aletheia: a space that is demanding, agile and lucid — capable of experimenting, evaluating and training the skills needed to learn and teach in the age of artificial intelligence.

What's already there

• The "My AI training" course — complete, from the "Why this course" preamble (Part 0) to the ethical and political stakes (Part 4), conclusion included. Each module mixes text, diagrams, expandable technical notes and an end quiz (75% pass threshold), with a cross-cutting recap per part. A persistent outline always shows where you are. And — a major new step — the entire course, quizzes and glossary are now TRILINGUAL (French, English, Chinese).
• Iris, the AI assistant — a teaching presence available in every lesson ("Ask Iris"): she answers with the module's content in context, never handing over the quiz answers. She also runs an AI news watch.
• Cheating or not? — 25 real classroom situations, shown one card at a time; the learner judges each, the score is weighted by severity, and an LLM mentor writes personalised feedback.
• Métis — the prompt sandbox — you write a prompt, it's evaluated on pedagogical quality and eco-responsibility, with concrete advice.
• Compass & comparison grid — an overview of models by bloc (US, China, Europe) and a comparison grid, fed live by public benchmarks (Artificial Analysis).
• The Mind Forge — brain training — while you learn to master AI, you train your own faculties: flash memory, mental arithmetic, N-back. A projector mode lets a whole class play together.
• Certification (beta) — a proctored, full-screen exam (tab-switch detection), opened by the teacher. Server-side grading, a pass threshold, matching interactions included.
• Classes & cohorts — a teacher creates a class, students join with a (time-limited) code, and the teacher tracks progress student by student; trainers run training cohorts.

AIQ — the mastery index

The AIQ sums up a learner's mastery of AI. It combines three dimensions: knowledge (quiz average, 40%), prompt practice (Métis evaluations, 35%) and course progression (validated modules, 25%). It is the foundation of the upcoming certification.

For whom?

• Students — middle school, high school, higher education — learning to learn with AI.
• Teachers — rethinking what we teach and how we assess.
• Free learners — curious adults who want to understand AI for their work or general culture.

Under the hood

A modern web app: Next.js 16 · React 19 · TypeScript · Tailwind CSS · Supabase (auth + PostgreSQL) · OpenRouter (multi-LLM orchestration). Hosted at Hostinger, in Europe. Trilingual for both the interface AND the course content. Installable as an app (PWA). Privacy: prompts pass through OpenRouter without the user's identity, are never used to retrain a model, and quiz/exam grading happens server-side. See the legal notice and privacy policy.

Roadmap

• Certification — higher levels — and a supervision view for the teacher (results and signals per student).
• Datagrade — integration as an evaluation and certification engine.
• LLM Label — a public "Aletheia Pedagogical Score" rating models for educational use.
• Mind Forge — next — long-term memorisation (spaced repetition) and new workshops.
• ★ Teacher modules — finalising the modules dedicated to the teaching profession (then translation).