The EU AI Act (2024)
The EU Artificial Intelligence Act (Regulation (EU) 2024/1689) informs how AI can be used across Europe, including in Irish schools. It is not a “schools’ policy” document, but it is relevant to the education sector.
It sets expectations about what is unacceptable, what must be transparent and what needs stronger safeguards (where AI could affect learners’ rights or outcomes).
A summary of the EU AI Act as it relates to schools
Schools that use AI tools are “deployers” under the Act, so they have legal responsibilities as users of AI, not just the companies that build AI. The Act is being introduced in phases, and one expectation for schools, is that staff who use AI should have a sufficient level of AI literacy, appropriate to their role and to the learners affected.
AI is already being used for a wide range of purposes in schools from lesson planning to differentiation of content to AI tools embedded in learning platforms. Leadership therefore needs clear boundaries for staff outlining which tools are approved, what is prohibited, and what requires approval, especially where decisions about learners could be influenced.
Most classroom practice would be deemed as “low-risk” if AI is used in a “drafting and support” role. It is always important that the teacher checks for accuracy, bias, tone and curriculum alignment. Where teachers use AI to draft rubrics or success criteria, the rubric should only support teacher judgement, it does not replace it. Teachers remain responsible for assessment decisions and recorded achievement.
The Act treats some education uses as “high-risk”, including AI intended to determine access or assignment (admissions or placement), evaluate learning outcomes, determine the level of education a learner will receive, or monitor behaviour during tests. Even if schools are not using AI in these ways today, these are the scenarios that would imply greater examination and consideration because of the potential impact on learners is much higher. The Act prohibits AI systems used to infer emotions in education institutions, except for medical or safety reasons, so tools claiming to detect pupil emotion, attention, stress or engagement from faces or voices should be prohibited for ordinary teaching and learning. For communications, a safe approach is to not use AI to generate or edit images, audio, or video of pupils and do not publish school media that could be mistaken as real.
The 4 P Approach to AI Implementation
For school leaders, the practical implications are best considered through the lens of the Irish AI Guidance which recommends a ‘4P Approach’ (Purpose, Planning, Policy Practice).
Start with baseline of how AI is being used and for what purpose. Developing AI literacy for staff, including what data must never be shared and safe use cases in the school’s context. Include an “AI tools and uses” section in the Digital Learning Plan, update the Acceptable Use Policy and connect AI decisions to GDPR and data protection processes. Finally, set clear positions on the tool categories schools are already using.
