Following a commitment made at the International Summit on the Teaching Profession in Reykjavik in March 2025, and backed by a resolution at EIS AGM, the EIS has worked with Scottish Government and other partners to develop guidelines and guardrails on AI in education.

The result is a framework built on five core principles — a framework that puts teachers, learners and human-centred education at its heart.

Increasingly, every one of us engage to some extent with Artificial Intelligence (AI) in our daily lives: it is embedded in almost every mobile device and online application we use. Many of the services upon which we rely – such as shopping, communications, transport and health – utilise AI to some degree.

The release of ChatGPT in late 2022 – followed by a stream of similar generative AI chatbots – propelled AI to the forefront of public consciousness.

A wave of hype around the transformative possibilities of AI ensued – not least in education, where firms marketing AI tools argued they could be the “magic bullet” to solve the teacher workload crisis: AI would write lesson plans, mark assessments, write reports and provide pupils with individualised mentoring and learning support!

However, back in the cold reality of Scottish education, EIS members began to raise significant concerns regarding pupils submitting AI-generated work as their own – particularly in coursework for national courses – which undermined the integrity of qualifications and placed additional pressure on secondary teachers to authenticate candidate work prior to submission to SQA.

EIS members in some areas also reported being pressured to experiment with AI tools in their teaching without adequate support.

More widely, members expressed worry as evidence emerged that unregulated generative AI presented significant safety and equity risks to children and young people and could adversely impact their ability to learn.

The EIS was clear that this technology – promoted by large-scale corporate interests and celebrated by government as the latest driver of economic success – was being driven into schools faster than its risks could be assessed. We argued from the outset that, whilst the potential of AI in education merited thorough exploration, national guidance and ethical guardrails – developed with teachers – was essential prior to schools engaging with AI.

This position found an echo in Professor Louise Hayward’s Independent Review of Qualifications and Assessment and in Scotland’s National Discussion on Education which asserted that education must remain a fundamentally human-centred endeavour.

Until now, AI in schools has unfolded in a largely unregulated and piecemeal manner – what some EIS members have described as a “wild west” approach. The Scottish Government’s Artificial Intelligence in Schools: Guidelines and Guardrails, published on 25 March 2026, are the result of sustained pressure by the EIS, and represent a framework for regulating AI use in schools that is underpinned by principles that the union has championed consistently: that teacher professional judgement is fundamental, and that teachers must maintain meaningful control over AI systems which impact their professional roles.

Further, the EIS argued strongly and successfully for the guardrails to be buttressed by a clear definition of ethical AI which encompasses fairness, transparency, accountability, safety, data privacy and human oversight.

AI use must ensure the safety and privacy of learners and staff by prioritising children’s rights, teacher judgement, ethical standards, and data privacy

The safety and privacy of learners and staff are non-negotiable. Local authorities are responsible for the AI systems used, and governance policies must be developed with teachers and their trade unions.

Within that framework, teachers can exercise their professional judgement in determining if and when to use AI. Schools require support to ensure that everyone who learns and works there understands AI risks and how to address them.

The use of AI must be underpinned by equity and fairness

AI must take account of the existing digital divide and actively address, rather than reinforce, inequality. AI apps themselves are not equal, with basic versions freely available and higher spec options available for purchase.

Many AI apps are marketed as particularly suitable for the differentiation of lessons and for learners who require additional support, but the role of the teacher in ensuring appropriate needs-led education, and the voice of individual learners, must be undiminished.

AI must support the aims of the Curriculum

AI should only be used where it supports learners to become: successful learners; confident individuals; effective contributors; responsible citizens. AI-use must not deskill them or disadvantage them in any way.

Digital literacy – including how AI works and how to use it responsibly across learning – is an essential skill for children and young people, but schools must maintain an appropriate balance between digital and non-digital learning.

AI must foster human connection and inclusivity

Human relationships are the core of education. AI must not replace teachers and other staff, or interactions between pupils and between pupils and staff. Nor should it be allowed to make decisions in schools. AI-generated outputs are subject to bias, inaccuracy and are not in themselves intelligent. Critical human oversight of AI is essential.

AI must support teachers

Teachers must be empowered to decide whether, when and how to use AI in their classrooms and should be supported by professional learning. AI must serve educational goals and not define them. It must never be used to monitor or evaluate teacher performance.

Whilst the publication of the guardrails is a welcome step forward, there is still much work to be done. Generative AI-related malpractice continues to be a major headache for secondary teachers, and the EIS will continue to work with Qualifications Scotland and other partners to address this.

The EIS nationally has a responsibility to support its local associations and school reps with advice and guidance to ensure the national guidance and guardrails are implemented effectively at local level and are adhered to by employers and management. Professional Learning – both in the ethical use of AI and in supporting the AI literacy of children and young people – will likely be an increasing priority for teachers in the near to medium future.

The EIS is committed to advocating for appropriate PL for teachers that is accessible and fully-resourced and is currently considering plans to include this in its own Professional Learning offer to members.