This is an official ODERSA website. Here’s how you know

The official domain

The address of this site ends in odersa.org. Every service the association runs sits on a subdomain of odersa.org and nowhere else. If the address in your browser’s address bar does not end in odersa.org, this site is not ours.

Free, and no account

Everything is open straight away. No sign-up, no account, no password, no subscription, no advertising. Nothing is held back for those who pay, because there is nothing to pay for.

No data collected

This site does not follow you: no tracker, no tracking cookie, no measurement tool built into these pages, and nothing measured on your device. Our host counts requests in aggregate, as any server that answers does: a total, never a profile. You do not have to take our word for it: open your browser’s developer tools, go to the Network tab, and reload the page. You will see the full list of what the site asks for. Everything comes from odersa.org, nothing goes anywhere else.

Free to reuse

The content is published under the CC BY 4.0 licence. You may copy it, translate it, print it and pass it on, for your classes as much as for the people around you, on one condition only: credit ODERSA.

What the official frameworks ask people to know

These frameworks describe expected competencies. None of them provides a course for reaching them. This page says what each one asks people to know, and where this course gives it.

A funder, a head teacher or a ministry often asks the same question: what does The AI Manual rely on? This page answers with public documents, each one opened at its address at the time of writing. It claims to invent nothing: it ties what those documents ask people to know to the precise place where this course gives it.

An expected competency, not a course

The four frameworks that follow describe competencies: what a person should be able to do when faced with an artificial intelligence system, at what level, at what age or in what job. None of them explains how a model builds its answer, why it makes things up with total confidence, or where its biases come from. That is no failing on their part: describing an expected competency and teaching how to reach it are two different trades. This course does the second, chapter by chapter, and shows it below, framework by framework.

The OECD and European Commission framework

The OECD and the European Commission, with the support of CodeAI, publish and launch this framework on 17 and 18 June 2026 in Brussels, after an open international consultation on a draft from May 2025 that drew more than two thousand contributors. It addresses primary and secondary education, and feeds the innovative domain of the PISA 2029 survey.

It describes four competence areas, nineteen competencies in all, each made of a piece of knowledge, a skill and an attitude: engaging with artificial intelligence, creating with it, managing it, designing it. Its area for managing AI calls for knowing what you choose to hand to a system, and what becomes of it.

What that framework allows anyone to state, this course gives in the chapter what you give it: the four distinct things that happen to a text, an image or a file handed to a service.

The UNESCO framework for students

UNESCO publishes this framework on 8 August 2024, during its Digital Learning Week. It addresses students in school education.

It describes twelve competencies across four dimensions, at three levels of progression: understand, apply, create. Its four dimensions are a human-centred mindset, the ethics of artificial intelligence, its techniques and applications, and the design of artificial intelligence systems. Its dimension on the ethics of AI calls for knowing why a system reproduces imbalances it did not choose.

What that dimension asks for, this course gives in the chapter where the biases come from: a demonstration where you skew the make-up of a corpus yourself, and watch the model give back exactly the skew you set.

The UNESCO framework for teachers

UNESCO publishes this framework on the same day, 8 August 2024: the first global framework of its kind designed for teachers themselves, and not for their students.

It describes fifteen competencies across five dimensions, at three levels: acquire, deepen, create. Its five dimensions are a mindset, the ethics of artificial intelligence, its foundations, its pedagogy, and professional development. Its dimension on a human-centred mindset calls for being able to recognise the situations where a machine’s answer has no place, not only those where it has one.

What that dimension expects, this course gives in the chapter when not to use it: three tests for deciding whether a situation allows anyone to rely on a machine’s answer, and why a detector never proves anything.

A non-European framework: AI4K12

AI4K12 is not a European framework, and that is exactly why it appears here: the work of artificial intelligence literacy belongs to no one continent. The AAAI and the CSTA have been publishing it since 2018, with support from the United States National Science Foundation, and have kept it up to date since. It addresses United States classrooms by grade band, from kindergarten to the final year of high school (K-2, 3-5, 6-8, 9-12 in its own terms).

It describes five big ideas: perception, representation and reasoning, learning, natural interaction, and societal impact. Its big idea of perception is about the way a system captures and represents an image, a sound or a text.

This course turns that question around: how to recognise that an image has been fabricated, not merely perceived. What that framework allows anyone to state, this course gives in the chapter a fabricated image: two images fabricated on the device that opens the page, their spectra side by side, and the limit of the method, since a little noise is enough to wipe it out.

What the frameworks ask for, and where this course gives it

The table keeps one point per framework, and stops at the chapters already published: a row with no real chapter behind it is not written.

An expected competency, the framework that asks for it, the page of the course that covers it
Expected competency Framework that asks for it Page of the course that covers it
Knowing what you choose to hand to a system, and what becomes of it OECD and European Commission, area for managing AI What you give it
Understanding why a model reproduces the imbalances of its training data UNESCO, framework for students, dimension on the ethics of AI Where the biases come from
Deciding whether a situation allows anyone to rely on a machine’s answer UNESCO, framework for teachers, dimension on a human-centred mindset When not to use it
Recognising the clues, and their limits, of an image fabricated by a machine AI4K12, big idea of perception A fabricated image

What is regulatory does not belong to this core

European Union law sets, on top of these frameworks, a separate obligation: its regulation on artificial intelligence requires every provider and every deployer of a system to support the artificial intelligence literacy of their staff, whatever the risk level of the system. That obligation has applied since February 2025.

It has nothing to do with the four frameworks above: a framework describes a competency to be reached, a legal obligation imposes an effort on an organisation, under an authority that can penalise it. The mechanism of a model is the same everywhere: a chapter describing it holds for every language of this course. A legal obligation does not: it holds for one territory and not for another. That is why it is dealt with apart, as a module placed beside this core, never inside this core: see the European obligation.

See also

What this site is, and what it commits to, is described on the about page. Every source of this course, framework by framework and theme by theme, is listed on the credits page.

Sources