AI in international schools: practical policy and classroom guidance for school leaders

AI has entered the school landscape at such high speed that school leaders across the world are now rapidly working to adapt to and regulate for the emerging technology.

The question for school leaders is: what should schools do about AI? There is much debate about what the best approach is – allowing, limiting or completely banning AI.

While there is no simple answer, this article explores how school leaders can create a practical and balanced AI policy.

Concerns for schools and AI 

UNESCO’s Guidance for generative AI in education and research lists five high-level concerns to pay particular attention to when evaluating AI use in schools:

1. Access and equity 

2. Human connection 

3. Human intellectual development 

4. Psychological impact 

5. Hidden bias and discrimination 

With this in mind, schools need to find the balance that works for them. This typically means incorporating AI into some, but not all, of their processes and operations, in and out of the classroom.

Doing this responsibly involves firstly deciding to what extent and for what purpose AI will be permitted. Implementing four tiers of permitted activity keeps it simple:

• Allowed, e.g. administrative drafting, initial lesson idea generation, or brainstorming starting points 

• Limited, e.g. teachers developing learning resources, students revising learning materials with AI operating as a virtual tutor 

• Supervised, e.g. drafting report card comments, initial feedback, or marking assistance that is strictly reviewed and verified by the teacher 

• Prohibited, e.g. students completing assigned work to pass off as their own, or uploading unanonymised student data

Across all four categories, apply a practical four-step test. Ask: 

1. What is the learning purpose? 

2. What student data protection is in place? 

3. What impact would using AI have on student thinking? 

4. What layer of human verification is involved?

How is AI being used in schools now? 

The possibilities of AI use in schools are far and wide. 

• For teachers: lesson planning, differentiation and learning resource creation are the most common uses. 

• For school leaders and administrative staff: tasks include reporting and data analysis. 

• For students: brainstorming and research, self-learning, and completing assigned work are the most frequent academic applications. 

The collective goal should be to put more time back into teaching and improving student outcomes. With that as the guiding beacon, focus stays sharp. 

Where AI can add value for teaching and learning 

With the right processes in place, teachers utilising AI can deliver meaningful value in the classroom. The caveat is the involvement of the irreplaceable human expertise that only real educators can provide.

Responsible and practical AI usage in schools should focus on augmenting, not replacing, teaching staff.

No AI lesson plan can replace professional judgement. The OECD's 2026 Reimagining Teaching in an Accelerating World report notes that teachers can use AI-generated ideas, warm-up activities or adapted materials as starting points – provided they check them for accuracy, curriculum alignment and fit for their specific students’ context before they reach the classroom. 

The report also notes that, when used wisely, AI tools can support conversational tutoring, personalised explanations, teaching adaptation, learning-data analysis and broader access for second-language learners, rural students and students who have historically been underserved. 

In schools, that opportunity is strongest when AI is used to support the cognitive struggle of learning rather than remove it, and when teachers retain judgement and oversight. 

The risks schools need to manage before expanding AI use 

Expanding school use of AI is not without risk. 

In their 2026 report, the OECD specifically calls out the risk AI poses of pulling students out of “the zone where real learning happens, when struggle, effort and growth take place”. 

The report also notes a risk for teachers who, by outsourcing too much of common processes, may end up hollowing out the value of their skill set. 

However, the associated risks of AI can be reduced with proper planning and governance. Key areas to minimise risk include: 

• Maintaining academic integrity: avoiding intentional and unintentional plagiarism by establishing clear parameters for student work 

• Critiquing AI outputs: teaching students to evaluate generated content, focusing on identifying biases, fabrications, and inaccuracies 

• Preserving cognitive growth: avoiding a reliance on AI which may lead to weakened creativity, problem-solving and critical thinking skills 

• Protecting data privacy: explicitly banning the upload of sensitive student data, names, or identifiable information into public or unvetted commercial AI models 

• Child safety and digital harm: aligning with regional and school safeguarding policies to ensure awareness of risks around image misuse, deepfakes and online safety 

• Ensuring staff readiness: providing regular, ongoing professional development so teachers feel confident guiding responsible use

What a practical AI policy for schools should include 

When developing a school AI policy, structure it around concrete guidance staff and students can actually follow: 

• Approved uses: clear examples of allowed low-risk tasks (e.g. lesson hooks, admin task support) 

• Restricted and supervised uses: tasks requiring explicit human oversight or specific tool parameters (e.g. feedback drafting, data summary) 

• Prohibited uses: non-negotiable boundaries (e.g. submitting AI work as original student output, uploading unanonymised data) 

• Assessment rules: clear guidelines on how assessment design shifts to measure process over final product 

• Disclosure expectations: explicit rules for when and how students and staff must cite or declare AI assistance 

• Data handling: strict privacy standards ensuring best practice and compliance with regional and school policies on data protection 

• Staff review: requirements for human verification on any AI-assisted output 

• Review frequency: scheduled updates to ensure policy keeps pace with rapid technological shifts 

• Parent communication: clear messaging on school expectations so families can support responsible device use at home 

Schools cannot control how students use AI all of the time, especially away from the classroom. As such, clear parent communication is critical: families should understand the school’s AI policy, expectations for student work, and the role they should play in supporting responsible device use. 

The way forward for AI in schools 

Like a lot of change management for schools, taking a phased approach works best. Start with low-risk use cases, provide training to staff, involve students and families, monitor the impact, and make regular updates to guidance documents and policies. 

So, should AI be allowed in schools? The answer is: schools need to find a happy medium that works for them. 

Most schools should neither ban AI entirely nor allow unrestricted use. Practically speaking, allowing, limiting or prohibiting AI use is dependent on the purpose, risk and oversight that can be managed. 

Schools should limit AI uses that replace the productive effort students need for learning, and instead allow uses that support thinking, feedback and access while keeping human judgement in the mix. 

Learn more about how you can prioritise informed decision making and create an environment where staff and students can thrive with our dedicated Advice hub, which covers all aspects of school management. 

School management Advice hub