European Commission and OECD set 19 AI literacy competences for schools

The non-binding framework covers how students engage with, create with, manage, and shape AI, with classroom examples for primary and secondary education.

The European Commission and OECD presented the AI Literacy Framework during the European Digital Education Hub flagship event in Brussels. Photo credit: Pia Ahrenkilde Hansen

The European Commission and the Organisation for Economic Co-operation and Development (OECD) have published a final AI Literacy Framework for primary and secondary education, establishing 19 competences intended to guide teaching, curriculum development, and education policy.

Presented at the European Digital Education Hub flagship event in Brussels, the framework was developed with support from CodeAI and an international expert group. The event brought together 150 policymakers, educators, researchers, and other education stakeholders.

Titled Empowering Learners for the Age of AI, the framework organizes AI literacy into four domains: engaging with AI, creating with AI, managing AI, and shaping AI. The 19 competences combine technical knowledge, human skills, and attitudes such as responsibility, reflection, curiosity, adaptability, and empathy.

The framework is intended for teachers, school and education system leaders, policymakers, learning designers, training providers, and families. It includes developmental expectations and classroom scenarios that can be adapted across subjects and education systems.

The publication is non-binding and is not intended to provide guidance on enforcing the European Union Artificial Intelligence Act. Its next scheduled use will include informing the OECD's first Programme for International Student Assessment (PISA) assessment of media and AI literacy in 2029, while a European Commission Education Package is due later in 2026.

Four domains move beyond the use of AI tools

The framework distinguishes AI literacy from simply knowing how to operate an AI product.

It defines AI literacy as the knowledge, skills, and attitudes required to understand how AI systems work, critically evaluate their outputs, and use them ethically and creatively.

Pia Ahrenkilde Hansen, Director-General for Education, Youth, Sport and Culture at the European Commission, wrote on LinkedIn: "Preparing young people for a world shaped by artificial intelligence starts in the classroom."

The Engage with AI domain focuses on identifying where AI is present, understanding its influence, checking outputs, recognizing bias, and considering environmental and ethical effects.

Students are expected to understand that AI systems are not human and do not possess awareness, intent, or authentic understanding. The framework advises educators to avoid language that incorrectly suggests AI "thinks" or "understands" in the same way as a person.

Students should also learn to verify AI-generated information using trusted sources and decide whether an output should be accepted, revised, or rejected.

The Create with AI domain covers the use of AI during brainstorming, design, feedback, and creative production. It asks students to maintain control of their ideas while considering originality, copyright, attribution, and intellectual property.

Manage AI focuses on deciding whether AI is suitable for a task, comparing different systems, allocating work between people and machines, and maintaining human oversight.

One classroom scenario asks students to separate the stages of writing an essay and determine which could be supported by AI and which require their own voice, reasoning, and judgment.

Shape AI introduces more technical work around evaluating and improving systems. Students may examine training data, test outputs against defined criteria, assess intended and unintended users, or suggest changes that make an AI system more inclusive.

Al Kingsley MBE, Chief Executive Officer of NetSupport and judge for the ETIH Innovation Awards, wrote on LinkedIn: "Rather than framing AI literacy as a purely technical skillset, the document places heavy emphasis on critical thinking, ethical awareness, learner agency, and social responsibility."

Classroom examples cover bias, misinformation, and human oversight

The framework provides basic, intermediate, and advanced learner expectations, although these stages are not tied to specific ages or grade levels.

Teachers are expected to select and adapt activities according to students' existing knowledge, local rules, subject requirements, and access to technology.

Examples include asking younger students to identify which everyday products use AI and comparing an AI-generated answer to a teacher's method for solving a mathematics problem.

At intermediate level, students could investigate whether AI-generated travel recommendations are outdated or fabricated, examine how facial recognition systems perform across demographic groups, or compare their own ideas with AI suggestions during a creative task.

Advanced activities include verifying AI-generated interpretations of historical events against primary and secondary sources, examining the effect of recommendation systems on public opinion, and evaluating how AI infrastructure affects energy and water use.

The framework also asks students to determine when AI should not be used. Its problem-solving guidance requires learners to compare AI-supported, human-led, and non-AI approaches before choosing a method.

AI ethics is integrated throughout the framework rather than treated as a separate topic. Students are expected to consider privacy, fairness, transparency, accountability, environmental impact, and who may benefit or be disadvantaged by an AI system.

The publication also addresses emotional reliance on AI. It warns that young people may treat AI companions as human-like sources of advice and encourages schools and families to challenge misconceptions that AI is all-knowing or capable of authentic relationships.

The framework cites research indicating that 88 percent of European teenagers aged 13 to 15 and 96 percent of those aged 16 to 18 use AI tools for learning and creative activities at least several times a week.

It also reports that 72 percent of teenagers surveyed in the United States had used AI companion tools, with some choosing to discuss serious matters with AI instead of another person.

Teacher support and international implementation remain central

The framework states that responsibility for AI literacy cannot sit with computing teachers alone.

Teachers across subjects are encouraged to identify where AI literacy fits within existing learning. Statistics, social science, computing, language, arts, and civics are among the subjects that could support different parts of the framework.

School leaders are expected to coordinate implementation, partnerships, and professional learning, while policymakers are asked to create conditions that support responsible adoption across education systems.

The publication cites the OECD's 2024 Teaching and Learning International Survey, which found that only one in three teachers used AI on average and three in four reported lacking the knowledge and skills needed to teach with it.

Almost 40 percent of teachers across OECD systems received AI-related training in 2024. The framework warns that adding AI literacy responsibilities without providing time and professional development could increase existing workload pressures.

It recommends that training move beyond product demonstrations and cover pedagogy, ethical reflection, student agency, evaluation, and the circumstances in which AI may weaken rather than support learning.

The final framework follows a draft published in May 2025. More than 2,000 people from over 100 countries contributed through surveys, written reviews, and focus groups, with teachers accounting for 41 percent of survey respondents.

Feedback resulted in greater emphasis on student agency, skepticism toward AI outputs, mental and emotional wellbeing, environmental costs, and developmental progressions.

The framework also states that evidence for generative AI's impact on learning remains mixed. Students may complete stronger work more quickly with AI support, but those results do not necessarily produce lasting learning gains.

The European Commission plans to publish its Education Package later in 2026, with digital and AI literacy included in its work on education system reform. The framework will also contribute to the PISA 2029 Media and Artificial Intelligence Literacy assessment, when AI literacy will be measured internationally for the first time.