A Standard Where There Was Only Improvisation
On June 24 the European Commission and the OECD published the final version of their AI literacy framework for primary and secondary education, presented to roughly 150 policymakers, educators, and researchers at a European Digital Education Hub event in Brussels. Titled Empowering Learners for the Age of AI, the document defines 19 competences that schools are expected to teach, organized into four domains: engaging with AI, creating with AI, managing AI, and shaping AI. It is the first time two bodies with this much convening power have agreed on a single map of what a student should actually know.
We read this as the end of the improvisation phase. For three years districts and ministries have written AI guidance in isolation, each reinventing the same definitions, and edtech vendors have marketed against a moving target. A shared taxonomy changes the procurement conversation. When a curriculum team can ask whether a tool supports managing AI or shaping AI as defined here, the market gets a common language, and that is usually the precondition for serious institutional buying rather than pilot purgatory.
What the Four Domains Actually Ask For
The four domains are deliberately more demanding than prompt writing. Engaging with AI covers identifying where AI is present, understanding its influence, checking outputs, and recognizing bias. Creating with AI asks students to use systems during brainstorming, design, and production while keeping human control. Managing AI is about deciding whether AI is even suitable and allocating tasks between people and machines. Shaping AI, the most ambitious, asks learners to evaluate and improve systems, examine training data, and assess inclusivity.
The 19 competences blend technical knowledge with human skills and attitudes, including responsibility, reflection, curiosity, adaptability, and empathy. Al Kingsley MBE, CEO of NetSupport, observed that the framework places heavy emphasis on critical thinking, ethical awareness, learner agency, and social responsibility. That framing matters for buyers: a tool that automates an essay does not satisfy a competence about maintaining human control, and ministries now have a document to point to when they reject the easy automation pitch.
The Data That Forced the Issue
The numbers behind the framework explain its urgency. Survey work cited in the launch found that 88 percent of European teenagers aged 13 to 15 use AI tools weekly for learning, a figure that climbs to 96 percent for those aged 16 to 18. On the other side of the desk, only about one in three teachers currently uses AI, three in four say they lack the knowledge to teach with it, and just 40 percent received any AI-related training in 2024. The students are already in, and the instruction has not caught up.
Pia Ahrenkilde Hansen, Director-General for Education, Youth, Sport and Culture at the European Commission, put the stakes plainly: Preparing young people for a world shaped by artificial intelligence starts in the classroom. The framework drew feedback from more than 2,000 contributors across over 100 countries, with teachers making up 41 percent of survey respondents. That breadth is what gives the document weight beyond Europe, even though it carries no force of law.
Non-Binding, but Not Toothless
It is worth being precise about what this is and is not. The framework is explicitly non-binding and is not intended to enforce the EU AI Act, which separately classifies AI used for student assessment and admissions as high-risk ahead of an August compliance deadline. Instead, the competences will feed PISA 2029, the first OECD international assessment to measure media and AI literacy, and a European Commission Education Package is due later in 2026 to build on it.
The PISA link is the quiet lever. Once a competence shows up in an international ranking, ministries that ignored it start treating it as a performance metric, because PISA scores move political careers. We expect the framework to harden in practice well before any regulation requires it, simply because measurement creates accountability. Vendors who align their roadmaps to these four domains now will be selling into a tailwind in two years rather than retrofitting later.
Why CTOs Outside Education Should Care
Enterprise leaders tend to file education policy under someone else's problem, and that is a mistake here. This framework is a forecast of the AI fluency baseline that graduates will carry into the workforce over the next decade. If the curriculum lands, the entry-level hire of 2030 arrives already trained to question model outputs, check for bias, and decide when not to use AI at all. That shifts the starting point for corporate onboarding and reduces the remedial AI training that learning and development teams currently shoulder.
There is also a more immediate read for anyone building or buying internal learning tools. The four-domain structure is a clean rubric for adult upskilling, not just school instruction. Managing AI, deciding suitability and allocating tasks between humans and machines, is precisely the judgment most enterprises say their workforce lacks. Borrowing this taxonomy for corporate programs gives organizations a defensible, externally validated structure rather than yet another bespoke competency model invented in a vacuum.
