Two European Commission reports released on September 22, 2026, found that four in ten Europeans aged 16 to 24 used generative AI tools for formal education in 2025, yet only 29.7% of EU teachers have received AI-related professional development. The reports warn that AI can raise performance without improving actual learning.
Key Takeaways
- Eurostat data cited by the Commission shows 40% of EU residents aged 16 to 24 used generative AI for formal education in 2025, with wide variation between countries.
- Only 29.7% of EU teachers surveyed in TALIS 2024 had taken part in professional development on AI for teaching and learning.
- Nearly two-thirds of European education systems now address AI through national strategies, guidance or policy frameworks.
- Around three-quarters of European education systems offer digital competence as a dedicated subject in lower-secondary school.
- Educational AI tools generally produce better learning outcomes than general-purpose chatbots, according to the Commission’s evidence review.
- The findings will feed into a forthcoming EU Education Package covering basic skills, teachers and digital infrastructure.
Two Reports Examine Generative AI and Digital Schooling Across Europe
The European Commission’s two publications approach the same question from different angles. The first, titled “Friend or foe? Evidence from the use of generative artificial intelligence in learning and teaching,” reviews scientific evidence and international studies on how generative AI affects students, teachers, parents and education systems. It covers cognitive development, critical thinking, self-regulated learning, equity and socio-emotional effects.
The second, the Eurydice report “Digital education at school in Europe 2026: Bridging gaps in access, teaching and learning,” maps how national systems are integrating digital technology, including AI, into schools. It draws on policy data from 38 education systems in the Eurydice Network and survey evidence from 24 systems that took part in the International Computer and Information Literacy Study (ICILS) 2023.
Together, the reports show that students have adopted generative AI faster than schools have adapted to it. That gap is the central finding for policymakers in Brussels and national capitals.
Better Grades Do Not Always Mean Better Learning
The Commission’s evidence review draws a distinction that matters well beyond Europe. Generative AI can improve student performance on tasks, but that does not necessarily translate into deeper learning. Students benefit most when AI supports their own thinking and they engage actively with the material, rather than handing the work to a chatbot.
The report raises a specific concern about over-reliance. When students use AI to complete tasks that are meant to build knowledge and cognitive skills, the tool can undercut the learning it appears to accelerate. The paper concludes that whether generative AI helps or harms depends on how it is used, and that its long-term effect on skills remains unclear.
The finding that purpose-built educational AI tools tend to outperform general-purpose assistants is likely to shape procurement decisions. Education ministries weighing whether to license dedicated learning platforms or rely on consumer tools now have an official EU evidence base leaning toward the former.
Teacher Training Emerges as the Weakest Link
The European Commission’s reports place teachers at the center of the challenge. AI can help educators prepare lesson plans, materials and assignments, but using it responsibly creates new professional demands. With fewer than a third of EU teachers trained in AI for teaching, most classrooms are adapting informally.
The Eurydice report reinforces the point. Digital competences increasingly appear in teacher education and professional development, but gaps persist in personalised learning, assessing online information and using AI. Both reports call for substantial initial and continuing training, backed by institutional guidelines on responsible use.
Executive Vice-President Roxana Mînzatu, who oversees social rights and skills, said students and teachers stand to gain from the new possibilities, but that risks around bias, privacy and over-reliance must be carefully addressed so these tools genuinely support learning.
Policies Are Spreading Faster Than Implementation
The Eurydice data shows that Europe has moved quickly on paper. Nearly two-thirds of education systems now have AI strategies, guidance or frameworks, and most support schools in drawing up digital plans. Yet only a minority systematically include digital education in quality assurance, meaning few systems measure whether those plans work.
Access is also uneven. Most systems have frameworks covering connectivity, hardware, software and technical support, but schools differ in infrastructure and access to advanced tools. Students from disadvantaged backgrounds may have fewer chances to build digital skills and can struggle at home without reliable devices or internet. The report notes that most countries have measures to address these barriers, but the divide persists.
What the Findings Mean Under the EU AI Act
The timing places European schools inside a new regulatory environment. The AI Office and national authorities began enforcing the EU AI Act on August 2, 2026, and the law lists certain education uses, such as systems that assess students or determine access to education, among its high-risk categories. Compliance timelines for some high-risk obligations were adjusted under the Digital Omnibus agreement reached in May 2026.
That leaves education authorities working on two tracks at once: building everyday AI literacy for teachers and students, while preparing for tighter rules on the specific AI systems that make consequential decisions about learners. The reports’ call for large-scale, long-term European research reflects how little hard evidence currently exists to guide either track.
For education systems outside the EU, the reports offer a useful benchmark. They document an adoption curve that many countries share, and they point to the same pressure point: teacher preparation, not student access, is where the transition is most likely to stall.
FAQs
What Did the European Commission’s Generative AI Education Report Find?
The report found that generative AI is widely used by young Europeans and can improve performance, but benefits depend on active use. It warned that over-reliance could weaken critical thinking skills.
How Many Young Europeans Use AI for School?
Eurostat data cited by the Commission shows four in ten people aged 16 to 24 in the EU used generative AI tools for formal education in 2025.
Are European Teachers Trained to Use AI?
Only 29.7% of EU teachers surveyed in TALIS 2024 had participated in professional development related to AI for teaching and learning.
How Many European Countries Have AI Education Policies?
Nearly two-thirds of European education systems address AI through national strategies, guidance or policy frameworks, according to the Eurydice report.
