Croatia: national framework for AI use in education

Benefits recognised, risks identified

The framework recognises AI’s potential to improve teaching and learning through:

• personalised learning;
• automation of teachers’ administrative tasks;
• learning analytics and early warning systems on student progress;
• assistive technologies that improve accessibility and inclusion.

At the same time, it identifies several risks linked to AI use in education, including:

• lack of transparency in AI decision-making;
• privacy risks arising from large-scale data collection;
• bias and unfairness in algorithmic outputs;
• academic dishonesty facilitated by generative AI tools;
• negative effects on student wellbeing and learning, including overreliance on technology and reduced critical thinking.

Key recommendations for AI in education

To address these risks, the framework emphasises that AI use in education must remain critical, supervised and ethically guided. AI should only be used where it clearly adds educational value, including:

• supporting lesson planning and differentiated instruction;
• enhancing project-based, research-based and creative learning
• providing feedback and practice opportunities for basic skills
• supporting inclusive education through tools such as speech-to-text systems.

In assessment, AI may assist teachers by generating tests, providing feedback, and offering learning analytics. However, final evaluation must remain under teacher control to ensure fairness, accountability, and the avoidance of bias caused by over-automated grading. 

The framework highlights teacher competence as a key condition for meaningful AI integration in teaching and learning. Students are expected to learn to use AI responsibly, transparently and ethically, while school leaders should ensure safe, equitable and lawful implementation. Policymakers are called to provide an appropriate regulatory framework and data protection safeguards, and parents should receive guidance to support informed participation and understanding.

The framework concludes that AI is a powerful educational tool whose value depends on human judgement, ethical application and clear pedagogical purpose. Trust, teacher–student relationships and critical thinking remain central to education, and AI should support – not replace – human teaching, learning and decision-making.

Impact of digital technologies on student wellbeing

Alongside the framework, a pilot survey on the impact of digital technologies on student wellbeing shows that AI tools are already widely used in Croatian schools. Around 40–50% of secondary school students report using AI daily for learning and homework. Teachers use AI less frequently, mainly for lesson preparation, and underline the need for more training, better access to tools and clearer guidance.

Although students generally view digital technologies positively, survey respondents express concerns about reliability, superficial learning, reduced concentration, mental health impacts and declining critical thinking. Recommendations therefore stress balanced, ethical and transparent AI use, with strong attention to privacy, data protection, cyberbullying prevention, critical thinking and maintaining a healthy balance between digital and offline life.

The study was conducted in 2025 in 46 schools (24 primary schools and 22 secondary schools), including VET schools. Questionnaires involved 3 913 students, 924 parents/guardians, and 382 teachers. In addition, focus groups were held with learners, parents, teachers, professional associates, and principals.

BrAIn: AI-based digital technologies in education

Alongside pedagogical guidelines and surveys, the BrAIn project pilots and gradually mainstreams AI curricula in Croatian schools. Additional curricula have been developed so that VET learners can take elective AI courses in all four years of secondary education. These curricula aim to strengthen learners’ digital and AI competences, critical thinking, ethical awareness and personalised, inclusive learning. The current focus is on supporting teachers with educational resources, training and a community of practice to facilitate them prepare and deliver AI classes.

The guidelines suggest possible future priorities for AI in education in Croatia:

• expanding personalised learning approaches;
• integrating smart recommendation systems;
• strengthening AI literacy;
• ensuring ethical use while limiting overuse;
• continuous professional development for teachers.

Read more

• Croatia: introducing AI in VET