AI in education: How to develop an AI policy for your school

AI in education is a reality. In a recent US survey, over 50% of students and teachers admitted to using AI. However, only 45% of principals could point to clear guidance for school leaders, students, or teachers.

This gap means staff and students are experimenting with AI without boundaries or safeguards, putting teaching quality, learning development and your school’s reputation at risk.

A clear AI policy allows you to take advantage of AI’s potential while protecting student development and confidential data.

How to develop an AI policy for teachers and students

If you don't already have one, the first step is to decide whether you need a standalone AI policy or can weave guidance into your existing assessment, safeguarding and academic policies.

For a standalone policy, think about how to structure it. For example, you might start with a high-level statement of principles. Then, follow this with use cases and guidelines tailored to each group (students, teachers, administrators, parents). This ensures everyone knows how to apply the rules consistently.

Establish a working group that includes school leaders, teachers, governors, parents and students. This will provide insights into how each group uses AI and what is and isn’t currently working in schools. Then decide who is responsible for developing your policy, overseeing implementation and keeping it updated.

What to include in your AI policy

AI evolves fast, and every school is different, so these are general guidelines. Remember, your policy must cover more than text-generation tools. Image generators, live voice translation, deepfakes and AI tutors are all widely used and should be included in your policy.

Here are our guidelines for what your policy should cover:

1. Student use of AI

This is a complex area for two reasons. First, young people usually lead when it comes to tech adoption, so teachers may be unaware of all the ways students use it. Second, students are still developing key skills like creativity, critical thinking and independent thought. Misuse or overreliance on AI puts all these at risk.

With this in mind, AI should be restricted to tasks like research, brainstorming, drafting ideas and project work. It is never a substitute for students’ initiative or creativity and should not be used for final drafts, exams or summative assessments.

It’s also essential that students can do something themselves before turning to AI. Otherwise, they won’t have developed the skills, in-depth knowledge and critical judgement needed to evaluate AI output for quality and accuracy. Even after mastering key skills, they should keep using them to avoid losing them. This is essential future-proofing that will help them stand out in a jobs market disrupted by AI.

Include recommendations for training on digital literacy, fact-checking, reliable sources and AI bias, limitations and hallucinations. Students must understand AI output isn't automatically true or safe and know how to evaluate it. Your policy should include guidance on student supervision, how to avoid inappropriate or harmful content and how to report concerns to a teacher.

Read more on the big debate on AI in education for a deeper dive into the benefits and risks.

2. Staff use of AI

While adults’ skills are more established, they, too, should not become overreliant on AI or overestimate its capabilities.

Before using it, staff should ask themselves whether it’s the most efficient way to achieve their goals. AI is a tool, not a default, and works best for:

• Repetitive tasks that follow a clear pattern, like creating large batches of similar worksheets
• Supporting lesson planning
• Brainstorming ideas for activities
• Generating activity or classroom instructions
• Compiling large amounts of information that would otherwise take hours to synthesise

Encourage them to think of AI not as an expert but as a junior colleague whose work they brief and review. AI can help teachers do more, faster, with research, lesson planning, marking and administrative tasks. But it’s no substitute for their experience and in-depth knowledge of subjects and students. For this reason, junior members of staff who are still learning should use AI with particular caution.

AI is best avoided when:

• The task requires deep knowledge that isn't documented anywhere
• Personalising standard information
• Editing or repurposing AI outputs
• Nuance matters more than speed
• There is no human review planned before people see it

Provide a list of AI terminology and include recommendations for staff training and ongoing professional development to stay updated on AI’s evolving capabilities.

3. Academic integrity

Once you’ve established acceptable use for your staff and students, think about how they should acknowledge and attribute AI use and how you’ll deal with breaches or non-compliance. So everyone is clear, establish different levels of misuse, from minor mistakes like not removing a prompt from a lesson plan to serious offences like presenting AI work as your own.

Voluntary disclosure is important as AI detection tools are unreliable and can unfairly penalise some students. Teachers’ knowledge of student ability is more reliable than any detection tool. If malpractice is suspected, take an approach of educating and understanding, not penalising, for first offences. For example, ask the student to explain their sources, methods and decisions. Then, help them to understand where they went wrong and how to use better practices in future. Of course, disclosure must also apply to staff and leadership.

4. Data protection

AI tools store and learn from whatever information is provided to them, posing serious threats to data protection. Your policy should specify what information can and cannot be shared with them. For example, confidential records or personal information that could identify a student or staff member should be totally off-limits. This includes full names, contact details, performance data, assessment scripts and medical or safeguarding information.

Apart from this, address specific scenarios like whether staff need leadership approval to sign up for a new AI tool or upload a student's work to get AI feedback.

AI in education: Safeguard learning with a solid policy

AI in education is a fact of life these days, so a strong AI policy lets you take advantage of its potential while limiting the risks. As well as covering student and staff use, data protection and academic integrity, you’ll also need to build training for staff and students. School leaders must model good practices as you roll the policy out across the school. Also, since AI is evolving fast, plan to gather feedback and use that feedback to inform regular reviews and updates. This will ensure your policy is a living, useful document that continues to serve the needs of staff and students.

Further reading

Read more about how students can use AI chatbots safely, discover some creative use cases for AI in the classroom, and learn how AI can support exam preparation.

Watch The Big Think: Assessment Evolved: Redefining formative assessment in a GenAI era.