1.5 Current State of AI in Education

Take Home message

• The EU AI Act is the legal basis. The EU’s rulebook for AI is in force since August 2024 and phasing in through 2027. It applies not just tech companies, but also to schools.
• AI literacy is a duty, not a course to tick off. Schools must support staff in building AI literacy.

Several AI uses are banned outright. Among them: inferring students’ emotions from facial or vocal data, AI that exploits children’s vulnerabilities to distort their behaviour, and building facial-recognition databases or inferring traits like race or religion from biometric data. Risk depends on the decision, not the product. Ordinary writing, research, translation and lesson-planning tools are not high-risk simply because they are used at school. They become high-risk once they influence admission, learning pathways, exam outcomes or employment decisions. Not everything AI-generated needs a label. The disclosure duty mainly sits with the provider; teacher-reviewed material with human editorial control is exempt. High-risk AI needs a human who can overrule it. AI output is never the final decision. You must understand its limits and be able to disregard or override it. The AI Act sits on top of other law, it doesn’t replace it. National data protection, education, employment, child protection and copyright law continue to apply alongside it....

If you remember only one sentence from this page: UseAI AIis already present in educational contexts but is not regulated. Structured frameworks are needed to support youreducators' professional judgement,development notimproving replacetheir it—stricterknowledge rulesand applyletting asthem soonteach aswith AI helps make decisionsand about grades, educational pathways or people.AI.

Current Uses of AI in Educational Contexts

Nowadays, artificial intelligence (AI) is embedded in everyday educational practice, well beyond pilot projects. Common applications include adaptive learning platforms, intelligent tutoring systems, automated grading, and AI-assisted lesson planning for teachers (OECD, 2026). The most visible shift, however, is the direct use of general-purpose tools by students and teachers alike. According to OECD's TALIS survey, 37% of lower secondary teachers reported using AI in their work in 2024, and 57% said it helps them improve lesson plans — though 72% also expressed concern about students passing off AI-generated work as their own (OECD, 2026). The OECD cautions that better task performance does not always mean genuine learning, warning of a risk of "metacognitive laziness" when cognitive work is offloaded to AI without pedagogical guidance.

Current State of Teaching With AI

...

Current State of Teaching About AI

...

AI Usage Across Educational Levels

 

IdentifiedWhat Needsis missing?

Alongside institutional frameworks and adoption data, it is important to understand how these dynamics play out in practice. A series of workshops was conducted with education stakeholders — including teachers and school leaders — to gather first-hand perspectives on the current use of AI in teaching. These workshops surfaced a set of concrete needs and concerns expressed by educators, covering practical, pedagogical, and training-related aspects. The following sections outline these needs in detail.

Understanding How AI Works. Both students and teachers need to understand how language models are built and trained, and what happens to their data once provided. They also need to grasp the nature of these models as prediction engines rather than databases, which explains their tendency to hallucinate. In light of this, teachers need a simple, formula-free, high-level course covering the basics of how LLMs work. Critical Thinking, Questioning and Verifying AI. Students and teachers need to be able to question AI-generated answers, and to recognise fabricated citations that appear credible but aren't. Practical verification skills are also needed, from checking AI output against official sources to spotting AI-generated images and videos. Students need to shift from following instructions to formulating them, thinking critically about the desired output. Ethics, Data Protection and Regulation. A general understanding of the ethical and legal aspects of AI use is needed, including the EU AI Act and GDPR, alongside practical rules such as never entering names or personal data into public AI tools. Safe and responsible use is essential to prevent illegal practices such as jailbreaking, along with awareness of related issues like bias in teaching materials and the unreliability of AI detectors. Finally, AI cannot be ignored or avoided, as doing so works against preparing students for its responsible use.

Participants emphasised the importance of a general understanding of the ethical and legal aspects of AI use, including the EU AI Act and the GDPR, together with practical rules such as never entering names or other personal data into public AI tools. Safe and responsible use was highlighted as essential to ensure that students do not engage in illegal practices, such as jailbreaking, while awareness was also raised around related issues including bias in teaching materials and the unreliability of AI detectors. Finally, participants stressed the importance of not refusing to engage with AI or pretending it does not exist, as avoidance was seen as counterproductive to preparing students for its responsible use.

Responsible Use

A further set of points focused on attitude and human agency, rather than on rigid rules. AI was seen as a tool to support, rather than replace, human work, to be used mindfully within a hybrid approach that leaves room for students' own thinking. Transparency was identified as a key principle in this respect, with students expected to indicate which parts of their work were their own. This approach was linked to the concept of "cyborg writing", a hybrid model in which humans and machines collaborate to draft, edit and refine text, treating generative AI as an extension of the writer's cognitive process rather than a substitute for it. Closely related to this is the principle of student ownership and accountability, whereby attributing a mistake to the AI is considered a failure of the student's own review process. Finally, participants noted that AI can mask personal weaknesses, a dynamic that may be either beneficial or problematic depending on the context. 

Practical Skills

Participants agreed that prompting and tool use should be treated as teachable skills, rather than something students are expected to acquire on their own. Both teachers and students were seen as needing explicit instruction in how to write effective prompts, while teachers additionally required training in prompt engineering to design lesson plans, differentiate skill levels and support grading. Prompting was also framed as a technical competence in its own right, closely linked to the ability to provide AI systems with the right information, for instance in debugging tasks. More broadly, participants highlighted the value of using AI as an assistant and co-creator, as well as the importance of being familiar with the wider landscape of available tools, beyond ChatGPT alone, in order to identify which tool is best suited to which subject.

Equity and Infrastructure

Concerns were also raised regarding equity and access. Financial disparities were identified as a key issue, with wealthier districts able to afford better and more expensive AI tools, thereby widening competence gaps between schools and institutions. To address this, participants called for access to AI tools to be guaranteed by states or national governments, with some expressing a preference for open models over corporate ones. The question of an appropriate minimum age for AI access was also raised, with reference made to Norway, where AI use is currently banned in elementary schools.

Assessment

Participants also discussed the need for a different approach to assessment. Rather than grading only the final product, emphasis was placed on assessing the process itself, together with students' reflections on AI-generated drafts. This includes AI-integrated assignments in which students are graded on their ability to critique, fact-check and edit AI output, an approach that raises the broader question of where the human role lies when teachers themselves also rely on AI for grading. Participants also stressed the importance of showing students AI's mistakes in a way appropriate to their age and level. A number of concrete classroom methods were proposed in this respect, including fact-checking AI-generated text, comparing responses from two different chatbots, debating the use of AI as part of homework assignments, and deliberately incorporating non-digital tools, such as flipcharts, into classroom activities.