Cutting Teacher Workload With AI: Where to Start

Workload is a long-running concern in schools. The Department for Education tracks it through its Working lives of teachers and leaders survey of teachers and leaders in state schools in England, which covers workload alongside wellbeing, flexible working and professional development. AI will not solve a workload problem on its own. Used carefully, though, it can take over some of the repetitive work that surrounds teaching, so that teachers have more time for the teaching itself.

This guide is for primary school leaders and teachers who want to get started without taking unnecessary risks. It is built around the DfE's guidance on generative artificial intelligence in education, which says generative AI has the potential to reduce administrative burdens and free up teachers' time to focus on teaching.

What the DfE says AI can help with

The DfE guidance lists several ways generative AI could support schools: creating educational resources, lesson and curriculum planning, tailored feedback and revision activities, and administrative tasks. Notice what those have in common. They are all jobs that produce a draft, a first version that a professional then checks and improves. That is the pattern to look for when choosing where to start.

The guidance is equally clear about the limits. AI-generated content can be inaccurate, biased, out of date or unreliable, and AI tools can produce false information presented as fact, which is known as hallucination. They cannot replace the judgement and deep subject knowledge of a teacher. The quality and content of any final document remains the responsibility of the professional who produced it.

Keep both halves of that in mind. AI is a fast drafting assistant, not a source of truth.

Start with low-risk, high-volume tasks

The best first uses involve no pupil data and no final decisions. They are the kind of task where a mistake is easy to spot and nothing reaches a child until a teacher has read it. Examples include:

• drafting a knowledge organiser for a topic you already know well;
• generating a bank of differentiated questions, then selecting and editing the best;
• rewriting a text at several reading levels as a starting point for adaptation;
• turning a lesson plan into a short, parent-friendly summary;
• drafting a first version of a routine letter or newsletter item; and
• producing a checklist or planning template for a recurring task.

In every case, the teacher reads and edits the output before using it. If a task would be risky to do without checking, it is not a good first use for AI.

Keep pupil data out of the tools

This is the rule that matters most. The DfE guidance says personal data should not be entered into generative AI tools, and that where it is strictly necessary, schools must make sure their use complies with data protection law. The government's guidance on data protection in schools explains the wider duties schools carry.

In practice, that means giving staff a clear, simple rule: do not paste names, assessment data, behaviour records, safeguarding information or anything else that identifies a pupil into a general-purpose AI tool. Staff can describe a situation in general terms instead. "A Year 3 pupil who finds subtraction with exchange difficult" gives an AI tool plenty to work with, without identifying anyone.

The DfE guidance also says schools should not allow pupils' original work to be used to train AI models without permission from the pupil or their parent or guardian. Teachers should therefore avoid uploading a child's writing to a tool unless the school has confirmed how that tool handles the content.

Agree the ground rules before the tools

A short, plain-English AI use policy prevents most of the problems schools later run into. It does not need to be long. Many schools find a single page enough to start with. It should answer a few questions:

1. Which tools are approved? A short list is easier to manage than a free-for-all.
2. What must never be entered? Pupil data and anything confidential should be named explicitly.
3. Who checks the output? The teacher who uses AI-generated material is responsible for its accuracy.
4. Who do staff ask when they are unsure? One named person removes hesitation.
5. How will pupils use AI, if at all? The DfE guidance says pupils need appropriate safeguards such as close supervision, and that schools should comply with the age restrictions set by AI providers.

The DfE also suggests schools review their homework policies in light of AI being widely available and consider how they engage with parents about AI tools. Both are easier to do once a basic policy exists.

Check the tool, not just the task

Before a school adopts any AI product, someone should ask whether it is designed for use with children and how it handles safety. The DfE has published generative AI product safety expectations, which set out what AI products used in education should meet to be considered safe. They are a useful reference when comparing tools or questioning a supplier.

Safeguarding duties do not change because AI is involved. The statutory guidance Keeping children safe in education continues to apply, and pupil use should be supervised.

Measure the time you get back

Enthusiasm is not evidence. If you want governors, colleagues or parents to take AI seriously, show them numbers. A simple method works well:

1. Pick two tasks that take a noticeable amount of staff time, such as preparing a weekly resource or writing routine communications.
2. Record how long each takes now, including the time spent checking and editing any AI-assisted draft.
3. Review after half a term. How much time did staff actually save, and did quality hold up?

Be honest about the checking time. A tool that produces a draft in seconds but needs twenty minutes of correction is not saving much. This kind of simple measurement also shows you which workflows deserve wider rollout and which to drop.

Build confidence in your staff, not just your policy

Written rules help, but confidence comes from practice. Staff who have tried AI tools on low-risk tasks, seen them make mistakes and learned how to check them are far better placed than staff who have only read a policy. The DfE guidance notes that early adopter research highlights the role of AI champions in building support among colleagues, so identifying one or two interested members of staff to help others is a sensible step.

Training does not have to be elaborate. A short hands-on session built around the tasks your staff actually do each week, with time to try things and ask questions, goes a long way.

Common mistakes to avoid

Most problems with AI in schools come from a small number of avoidable habits. These are the ones worth warning staff about.

Using AI output without reading it. The most common mistake is also the most serious. A fluent, confident answer is not necessarily a correct one. Every piece of AI-generated material needs a teacher's eyes on it before it is used, and anything factual should be checked against a reliable source.

Letting AI make decisions about pupils. AI can help draft a resource, but it should not decide anything about a child. Judgements about progress, behaviour or support belong with the professionals who know the pupil.

Trusting content that may not be original. The DfE guidance warns of intellectual property risks, including secondary infringement if AI outputs draw on unlicensed material. Treat generated text and images as drafts that you adapt, not as finished work to publish unchecked.

Using personal accounts and unapproved tools. If staff use whichever free tool they happen to find, the school has no control over where information goes. A short approved list, with clear instructions, is safer and easier.

Doing it all at once. Schools that try to change everything in one term tend to lose staff goodwill. Starting with two small tasks and building from there is slower on paper and faster in practice.

Forgetting to tell parents. The DfE recommends that schools consider how they engage with parents on AI tools. A short, calm explanation of what the school does and does not do with AI prevents rumour and worry later.

A simple starting plan

If you would like a concrete place to begin, this sequence works for most primary schools:

1. Agree a one-page AI use policy and name a person staff can ask.
2. Choose two low-risk tasks to try, with no pupil data involved.
3. Give staff a short, practical session on using AI as a drafting assistant and checking its output.
4. Record the time taken on those tasks before and after.
5. Review after half a term, then widen or adjust.

Each step is small. Together they give a school a safe, evidence-based way into AI without needing to commit to anything large.

Where to go from here

The details will differ from school to school, but the principles stay the same: keep pupil data out, keep teachers in control, check everything and measure what you gain. The DfE's published guidance is the best place to start, and it is worth reading in full.

If you would like help putting this into practice, you can read more about AI training for UK schools, or see how we support primary schools in East Sussex. You can also book a free 15-minute call to talk through where your school is starting from.