Artificial intelligence in European schools: experiences and challenges in Italia, Lithuania, Spain and Greece

In fact, however, many institutions are still hesitant about this novelty. "The guidelines also provide for data protection impact assessments and the protection of families or students who do not want to use AIDs," explains Benassi, "The school, in fact, becomes responsible for providing the service. It is a new field, there is a lack of established practices, and this generates fear'.

Since 2024, the Ministry has been conducting an experiment, which is still ongoing, involving several schools in Lombardy, Tuscany, Calabria and Lazio. The project is supported by Google and is aimed at personalising teaching.

In addition to the ministry's initiative, Indire is also developing a path for the incorporation of Ai in teaching that, as Benassi recounts, currently involves 26 schools: 'The goal is not only to help them comply with the guidelines, but to build a systemic vision for the use of artificial intelligence. We need to overcome small isolated experiments at the level of individual classrooms or teachers'.

In academia, the main coordinating authority is the Conference of Italian University Rectors (Crui). In 2025, Crui's ICT group signed an agreement with OpenAI, marking a decisive step towards the integration of ICT in the university system.

As of the academic year 2025/2026, institutes can more easily access ChatGPT Edu, a version of the OpenAI chatbot designed for responsible use by students, researchers, lecturers and university staff.

Thanks to this agreement, the first of its kind in Europe, licences are available on favourable terms for associated universities, and Italian research centres can also benefit from these advantages.

In Lithuania artificial intelligence enters the school system in a structured way through a framework of rules designed to balance innovation, responsibility and human control. The Ministry of Education, Science and Sport has published guidelines 'Schools in the Age of Artificial Intelligence (AI)', which provide each institution with criteria for defining internal rules on the use of AI in teaching and administration, without imposing formal obligations.

The document clarifies that schools may only adopt tools that comply with European legislation and the protection of personal data. Each institution is asked to draw up a list of permitted applications, approved by the school head, excluding technologies based on emotion recognition, biometric tracking or social rating systems. The risk assessment criteria are aligned to the AI Act and developed together with the Ministry of Economy and Innovation.

At the heart of the recommendations is the principle of added educational value: AI should only be used when it concretely improves learning, e.g. to personalise exercises, reduce repetitive tasks or support creativity and analysis. The decisions always remain with the teachers, who determine when to use AI and when to privilege students' autonomous work. It is also recommended that use by pupils should take place with teacher authorisation and that content generated with AI tools should be clearly declared.

In parallel, the support ecosystem is growing. The national education agency and the association EdTech Lithuania offer advice and a virtual assistant to help school leaders assess the safety and suitability of tools. Universities are involved in applied research projects on the impact of AI on learning and assessment, with the aim of defining specific skills for teachers and developing new educational platforms.

On the operational side, Lithuania has invested more than EUR 1.7 million in Microsoft 365 Copilot licences: free access for around 330 thousand students and more than 20 thousand teachers, accompanied by dedicated training programmes. According to OECD data and ministerial surveys, AI is already widely used in Lithuanian schools, with adoption levels above the European average.

In Spain, artificial intelligence is already widely present in school life, but it proceeds in the absence of a clear national regulatory framework. According to the report El impacto de la AI en la educación en España (Empantallados.com and GAD3), eight out of ten students have used AI tools to perform homework, integrate subjects or prepare for exams. Use does not only concern pupils: 73% of teachers and 69% of parents say they have used these tools, especially chatbots such as ChatGPT, which is a frequent source of information for 40% of students.

Despite its widespread use, schools are moving in short order. A Capgemini Research Institute survey of 1,800 secondary school teachers found that 48% of institutions have blocked or restricted the use of AI, 19% allow it for specific uses and 18% are still evaluating its applicability and usefulness. In the absence of agreed rules, some schools and universities have introduced tools to detect papers produced with AI in an attempt to protect academic integrity.

The debate remains open. Experts and researchers emphasise the transformative potential of generative AI, but warn against the risk of overuse that weakens critical thinking and human interaction. University studies, such as those of the University of Murcia, call for integrating these technologies as a support, not a substitute, for learning processes, promoting digital literacy and interdisciplinary training for teachers and students.

In Greece, artificial intelligence officially entered schools in the school year 2024-2025, as part of the government's 'Digital School' programme, funded by the Recovery and Resilience Facility. The initiative envisages the gradual integration of digital services for students, teachers and families, including educational tools with AI components. However, modalities and criteria for use remain unclear.

There is currently no specific national regulation on the use of AI in the classroom. Adoption is largely left to the initiative of individual teachers, who report increasing use by students, especially of chatbots such as ChatGPT for homework and research. According to Nektarios Kordis, a member of the board of the Federation of Secondary School Teachers (OLME), the absence of common guidelines generates ambiguity: 'Students rely on the algorithm to write essays, with standardised and superficial answers.

The government supports the need for international cooperation to define common rules on AI, an issue also addressed at the OECD by Education Minister Kyriakos Pierrakakis. But strong misgivings persist in the country. Part of the educational community fears that the digital programme could result in a reduction in teaching staff and denounces structural deficiencies: inadequate school buildings, slow connections and difficulties in maintaining basic IT equipment. In this context, AI appears as an opportunity yet to be defined, suspended between announced innovation and operational criticalities.

*This article is part of the European collaborative journalism project "Pulse" and was contributed by Justė Ancevičiūtė (Delphi, Lithuania), Ana Somavilla (El Confidencial, Spain), Eva Papadopoulou (EFSYN, Greece).