Abstract:

Artificial intelligence is already part of many students’ daily lives, but schools are still deciding how to respond. This paper presents findings from Estonia based on large-scale student data collected in 2024 and 2025 through three national cybersecurity competitions with 1-12th graders that also included questions about AI in education. In 2024, participation ranged from 3,826 to 8,036 students across 75 to 170 schools; in 2025, it remained similarly high, ranging from 3,638 to about 8,795 students across 96 to 196 schools.

Across age groups, students' message was clear: they do not want AI ignored, and most do not want it banned. Instead, they expect schools to address it openly, fairly, and in ways that support learning. Even among younger learners, AI is already present in school life. In the 2024 CyberPin data, about one-third of 5th–6th-grade students reported using AI both at school and at home, and nearly one-third used it to help with homework. Some students also admitted to using AI even when it had been forbidden, showing that AI use is already happening regardless of school policy.

The strongest expectation was the need for clear rules. In both 2024 and 2025, most students wanted schools to explain when AI use is allowed, what counts as fair use, and where the limits are. Students appear to see rules not mainly as restrictions, but as a way to make expectations transparent and consistent across subjects and teachers. This matters especially in situations where some teachers actively discuss AI while others avoid it or treat it mainly as a risk.

A second major expectation concerns practical teacher guidance. Students do not only want warnings about misuse. They want teachers to show how AI can be used meaningfully in real subjects and learning tasks. The findings suggest that many students receive more guidance about how AI should not be used than about how it could support learning. This creates uncertainty and leaves students to experiment on their own.

Students also want independent learning materials and, especially among older groups, opportunities to move beyond use toward understanding and creation. Many learners expressed interest in learning more about AI at their own pace, and a notable share also wanted to learn AI programming or gain deeper technical knowledge. This shows that students do not want to remain passive users of AI tools. They want schools to help them develop informed and active AI literacy.

At the same time, support for a complete ban on AI was low in both years. Students were generally more ready for guided integration than for prohibition. However, they were not uncritical. Older learners, in particular, raised concerns about AI in assessment, fairness, responsibility for mistakes, transparency, privacy, and the risk that AI may reward results without recognizing effort, growth, or creativity. This shows that students do not want schools to automate education, but to implement AI in a human-centered way where teacher judgement remains central.

Overall, the findings suggest that schools should respond through four connected steps: establish shared rules, provide practical subject-based guidance, offer self-study materials, and keep strong human oversight in teaching and assessment. Students are ready for schools to take an active role in AI education, but they expect that role to be clear, fair, and educationally meaningful rather than simply restrictive.

Keywords:

Artificial intelligence, K–12 education, Student Voice, AI Literacy, School Policy, Teacher Guidance, Assessment, Human-Centred Education.