Understanding and Using Artificial Intelligence in Primary and Secondary Schools of the Republic of Croatia

Authors: Lidija Kralj & Darija Dasović Šebalj, Suradnici u učenju (Partners in Learning)

Translation: ChatGPT-5.2 (OpenAI, 2026). Consolidation and final review: Arjana Blažić

Summary of the research with students

The research was conducted by the Partners in Learning association during November 2025 in the Republic of Croatia on a sample of 2,514 primary and secondary school students, the majority of whom were aged between 11 to 14 (72%). Most respondents attend schools in urban settings (cities and small towns), while 22% come from rural schools.

The results show that the use of artificial intelligence (AI) is present but not universal: around 60% of students use AI tools at least occasionally, and both frequency of use and confidence increase with age. The study indicates that AI use among students is widespread but not fully uniform. More than half of students use AI tools at least occasionally, with usage increasing significantly with age. Younger students rarely or never use these tools, whereas older students increasingly use AI regularly, indicating growing digital competence and independence with age. Differences by school location exist but are small and not of major importance: students in urban areas use AI tools only slightly more often than those in smaller communities.

The data indicate a moderately developed understanding of how artificial intelligence works. Almost half of the students believe they understand how AI makes decisions; however, a very large proportion of neutral responses suggests uncertainty or partial understanding. Students who use AI tools more frequently are significantly more likely to report understanding how they work, confirming that experience and practical use play a key role in developing understanding, particularly in younger age groups.

Levels of confidence in using AI show a similar pattern. Approximately one third of students feel confident, the largest group report neutral feelings, and one quarter do not feel confident. Confidence increases with age but remains limited by a high proportion of undecided responses. Students who believe that artificial intelligence will be important for their education and future careers also feel significantly more confident in using it. This represents the strongest relationship identified in the study and highlights the importance of meaningfully linking AI use to future educational and professional contexts.

Students’ attitudes towards the impact of artificial intelligence on learning are generally positive but marked by a high degree of indecision. Almost half of students believe that AI can make learning more engaging, while a large proportion remain neutral. This pattern suggests that students are not sceptical but rather lack clear examples and concrete educational experiences that demonstrate how AI can meaningfully contribute to learning.

Students demonstrate a high level of awareness of ethical challenges and recognise that AI can be inaccurate and biased. A large majority are aware that AI may provide incorrect or biased answers and believe it is important to verify the information it provides. This awareness increases with age and represents a strong foundation for developing responsible and critical use of artificial intelligence. Students who recognise the limitations of AI are also more likely to express a desire to learn more about safe and responsible use, confirming that critical understanding encourages motivation to learn rather than fear or rejection of the technology.

However, across many categories (understanding how AI works, confidence, importance for future careers), “neutral” attitudes prevail. This dominant neutrality does not indicate indifference, but rather a lack of experience, clarity, and confidence. This creates an exceptionally favourable context for the systematic introduction of artificial intelligence in education through the gradual development of AI literacy, guided activities, links with curriculum content and career scenarios, and strengthening students’ confidence. Such an approach has the potential to transform existing neutrality into knowledge, competences, and responsible, thoughtful application of artificial intelligence.

• Demographics and sample context: A total of N = 2,514 students participated in the study. The most represented age group was 11–12 years (45.4%; n = 1,141), followed by 13–14 years (26.6%; n = 668), while 7–10-year-olds accounted for 16% (n = 403). The groups under 7 (0.6%; n = 15) and over 18 (1.8%; n = 46) were very small and their results should be interpreted with caution. Most students attend school in a city (45%; n = 1,136), followed by a small town (32%; n = 815) and rural areas (22%; n = 563).
• Use of AI tools is widespread and increases with age: 59.3% of students use AI at least occasionally (sometimes, often, or always), and the proportion who use AI “often or always” rises from approximately 12% among those aged 7–10 to around 52–55% in secondary school groups.
• Understanding and confidence are linked to exposure: Students who always or often use AI are more likely to report understanding how AI makes decisions (~55% agreement), whereas those who rarely or never use it report such understanding far less frequently (~28–41%). Confidence is “neutral” for the majority (42%) but increases slightly with age.
• Attitudes are mostly positive but highly neutral: Approximately 48% believe AI can make learning more engaging; around 80% recognise that AI can be inaccurate or biased, and around 70% believe information provided by AI should be verified. Neutral responses remain high (approximately 30–45%).
• Associations: Students who believe AI will be important for education and careers show higher confidence in using AI (χ² = 580.75; V = 0.241; Spearman’s ρ = 0.336). This is the only relationship of moderate strength and is pedagogically the most important: belief in importance is associated with higher confidence.
• Most statistical associations are weak in strength (Cramér’s V ≈ 0.04–0.11). The strongest links are between age and attitudes (e.g. “I want to learn more about AI” V ≈ 0.106; “It is important to verify information” V ≈ 0.095). School location has even weaker effects (V ≈ 0.036–0.063).

Full text in English. Full text in Croatian. 

Infographic is made with Google NotebookLM

Summary of the research with teachers

The study conducted by the association Suradnici u učenju (Partners in Learning)  in November 2025 examined the level of understanding, frequency of use, self-confidence, attitudes, and ethical awareness of teachers regarding the application of artificial intelligence (AI) in primary and secondary schools in the Republic of Croatia. The study was carried out on a sample of 399 participants. Respondents were predominantly middle-aged (31-50 years), and schools from urban, suburban, and rural areas were almost evenly represented, enabling relevant analysis with respect to demographic and spatial factors.

The results show that AI is already widely present in teaching practice: as many as 94% of participants use AI tools at least occasionally, while 68.9% use them sometimes or often. This points to a high level of acceptance, but also to some unevenness in the integration of AI into the teaching process. Most participants (80.2%) believe they understand how AI makes decisions, while the level of open disagreement is very low. Differences in understanding by age and school location are small and statistically insignificant.

More than half of participants (54.9%) feel confident when using AI tools, but a substantial share (36.3%) takes a neutral position, indicating a need for additional support and professional development. Age appears to be one factor related to confidence: younger participants are more confident, while older age groups, especially in rural settings, more often express uncertainty. School location has no statistically significant effect.

In terms of attitudes and motivation, most respondents show a strong interest in exploring how AI works (70% positive responses) and believe that AI can make teaching more engaging (74% positive responses). A strong and statistically highly significant association was found between interest in exploring AI and a positive perception of its pedagogical value (Cramér’s V = 0.46; p < 0.001), confirming the importance of curiosity and understanding in shaping attitudes toward technology.

Participants also demonstrate an exceptionally high level of ethical awareness: 95% recognize that AI can produce inaccurate or biased information, and 97% highlight the importance of verifying information provided by AI. Furthermore, 86% of participants express a desire to learn more about the safe and responsible use of AI in education, and 79% believe that AI will play an important role in the future of education, with minimal differences by age and school location.

Overall, participants express positive attitudes toward the pedagogical potential of AI, believe it can make teaching more engaging, and consider it likely to play an important role in the future of education. At the same time, there is high awareness of AI limitations, including the possibility of inaccurate and biased information, and a strong consensus on the need to verify information. Interest in further learning about safe and responsible AI use is very high across all age groups and school locations.

In conclusion, the results indicate that teachers in Croatia demonstrate a high level of openness, critical thinking, and readiness for professional development in the field of AI. AI is not perceived as a threat, but as a potentially valuable pedagogical tool, alongside a clearly expressed need to further strengthen competences for its safe, ethical, and thoughtful application in the education system.

Full text in English. Full text in Croatian.  Additional infographics for presenting research.

Infographic is made with Google NotebookLM