Mail

When the internet arrived, people said "we need to teach search skills." When smartphones spread, they said "we need to teach digital etiquette." Now, as AI weaves itself into daily life, a new consensus is forming: "we need to teach people how to read, judge, and think critically about AI."

The European Commission and OECD have formally answered this call. The AI Literacy Framework for Primary & Secondary Education, whose draft was published in 2025, is heading toward its final release in 2026.

Why an AI Literacy Framework, and Why Now?
Who Built It, and How?
The 4 Domains and 22 Competences
What It Looks Like in the Classroom — Lesson Scenarios
The Connection to PISA 2029
What It Means for Education Systems

Why an AI Literacy Framework, and Why Now?

In 2026, 86% of higher education students globally are using AI as their primary research and brainstorming tool. In 2024, that figure was 66% — a jump of more than 20 percentage points in two years.

But using AI and understanding and critically evaluating AI are very different things. Not knowing why AI gives the answer it gives, what biases it might carry, or when it should not be trusted — using AI without that understanding is potentially more dangerous than using a calculator without understanding arithmetic.

In Europe, the EU AI Act is now fully in force, requiring AI systems to meet standards of transparency and accountability by law. In step with this legal context, the demand for schools to formally build AI literacy into curricula has grown significantly.

Who Built It, and How?

This framework was not built alone.

The European Commission and OECD jointly led the effort, with support from Code.org and education experts from around the world. After a review draft was published in May 2025, input was gathered from approximately 1,000 stakeholders — teachers, researchers, policy makers, and parents.

The final version is expected to launch in 2026, and it will not just be a document. It will include AI literacy exemplars — concrete classroom examples showing how teachers in science, social studies, language arts, and other subjects can integrate AI literacy into their teaching without needing specialized technology access.

The 4 Domains and 22 Competences

The framework organizes AI literacy into four broad domains, with a total of 22 specific competences across them.

1. Engaging with AI Using AI tools in practice and developing a foundational understanding of how AI works. Not just "knowing how to write good prompts" — but building experiential understanding of how AI responds and where it falls short.

2. Creating with AI Using AI as a collaborative tool for creative work, problem-solving, and projects. Writing, art, coding, data analysis — experiencing the process of building something together with AI.

3. Managing AI Critically evaluating AI-generated information and judging its reliability. Understanding AI bias, the possibility of error, and privacy implications — developing the ethical judgment to use AI responsibly.

4. Designing AI Understanding the principles behind how AI systems are built, and exploring basic AI model design or data training concepts. This domain is particularly relevant for middle and high school levels.

A key feature: this framework is explicitly not confined to computer science classes. It is designed to flow across all subjects — exploring AI bias through a science experiment, analyzing the trustworthiness of AI-generated text in a language arts class, debating AI ethics in social studies.

What It Looks Like in the Classroom — Lesson Scenarios

The framework draft includes scenario examples that make this concrete.

Primary school: AI Storywriter Experience Students suggest main characters to an AI, which then writes stories. If the AI keeps generating similar characters (blonde princes, thin protagonists), the class discusses why. Students discover where AI bias comes from — through storytelling, not technical instruction.

Middle school: Fake News Detective Agency Students receive a mix of AI-generated and human-written texts and try to tell them apart. Discussing what clues they used builds critical reading skills and AI literacy simultaneously.

High school: Build Your Own Classifier Students build a simple spam email classifier themselves. Experiencing how results shift depending on training data, they discover firsthand that AI is not an "absolute judge" but a "reflection of data."

The Connection to PISA 2029

Here is why this framework is more than a set of recommendations.

The OECD has confirmed that the 2029 PISA (Programme for International Student Assessment) will include a new domain: Media and AI Literacy (MAIL). This framework is the conceptual foundation for the PISA 2029 item development.

The implications are significant. PISA involves more than 80 countries. When AI literacy becomes part of what PISA measures, governments face real pressure to integrate it into national curricula. A recommendation becomes a de facto global standard.

What It Means for Education Systems

In many countries, the debate about AI in education has centered on how much to let students use AI. The EU-OECD framework asks a different question: how well can students understand and critically use AI?

Knowing how to use an AI tool is different from understanding AI — just as knowing how to drive is different from understanding how an engine works. And for road safety, a degree of mechanical understanding matters.

When students are measured on AI literacy in PISA 2029, it will reveal how well their education systems prepared them. That preparation needs to begin now.

Related Posts

How is AI literacy being taught in your school? Share your experience in the comments!

Sources:

The EU & OECD AI Literacy Framework 2026 — A Complete Guide

Table of Contents