1. What could AI offer schools?

AI tools can assist with some tasks that take teachers and support staff considerable time. They may help draft lesson plans, create practice questions, adapt explanations to different reading levels, generate activity ideas and organise routine administrative information.

When used carefully, these tools may give teachers more time to focus on teaching, feedback and individual pupils. AI can also help young people explore how digital systems work and develop the critical thinking skills needed to use them responsibly.

However, AI-generated material can be inaccurate, biased, unsuitable for a particular age group or inconsistent with the curriculum. Teachers still need to check resources, adapt them to their pupils and use professional judgement.

The Department for Education's guidance for England states that AI should support, not replace, the valuable relationship between teachers and pupils.

2. Public investment: what is being promised?

The UK Government has announced several initiatives intended to support the safe use of AI and technology in education in England.

AI-powered tutoring for disadvantaged pupils

In January 2026, the Government announced plans for AI tutoring tools that could support up to 450,000 disadvantaged pupils each year. The programme is intended to provide personalised help in subjects including English, maths, science and modern foreign languages. The tools are being developed with teachers and are intended to complement classroom teaching, not replace it.

In April 2026, the Government invited education technology companies and AI developers to help build and test these tools, with wider availability targeted from 2027 if the programme proves successful.

Support for teachers and school leaders

The Department for Education provides free materials for school and college staff on understanding AI, interacting with generative AI and planning safe implementation. These resources were updated for the 2026–27 academic year in May 2026.

Technology and AI opportunities

The Government's wider digital education plans also include improving access to technology, supporting staff and exploring tools that may help reduce workload and address attainment gaps.

These initiatives show that public investment is being directed towards AI in education. However, a funding announcement or programme target is not proof of completed delivery, improved attainment or long-term value for money. Those outcomes need to be measured and reported.

3. Does AI actually reduce teachers' workload?

There is growing evidence that AI can reduce the time required for certain preparation tasks, although the results depend on the tool and how teachers use it.

An Education Endowment Foundation evaluation published in October 2026 examined Oak National Academy's Aila lesson-planning assistant in a trial involving primary teachers. Teachers using Aila spent around 2.5 hours a week on lesson and resource preparation, compared with 3.3 hours for teachers in the comparison group. The evaluation found that resource quality was not worse in the Aila group.

The findings are encouraging, but there are important qualifications. Use of the tool was relatively limited and declined during the trial. Teachers had mixed views, and fewer than one in five felt that the tool had meaningfully influenced their teaching and planning.

A separate EEF trial involving 259 teachers across 68 secondary schools found that teachers using ChatGPT alongside a practical guide spent 56.2 minutes per week on relevant preparation, compared with 81.5 minutes in the comparison group.

These trials support the possibility of time savings for specific tasks. They do not establish that every AI tool will work equally well, that all schools will achieve the same results or that AI automatically improves pupils' learning.

Schools should evaluate whether time saved is genuinely redirected towards feedback, lesson quality, pupil support and teacher wellbeing.

4. Training must be part of the investment

Giving staff access to a tool without providing time and training can lead to poor results. Teachers need to understand how AI generates responses, why it can make mistakes and how to check its work.

School leaders also need to know how to assess proposed products, review data protection arrangements, manage safeguarding risks and decide whether a tool is appropriate for a particular task or age group.

Good training should cover:

• Recognising inaccurate, biased or fabricated AI outputs.
• Checking curriculum alignment and the suitability of learning materials.
• Protecting pupil information and following school policies.
• Understanding copyright, attribution and intellectual property.
• Using AI to support accessibility without making assumptions about pupils' needs.
• Explaining acceptable use to pupils and teaching them to question AI-generated information.
• Knowing when human expertise is required and when a tool should not be used.

Training also needs to be practical. Staff should have time to experiment safely, share experiences and raise concerns without pressure to adopt technology simply because it has been purchased.

5. SEND and disability inclusion: will every learner benefit?

AI may help some pupils access information through simplified explanations, alternative formats, translation, structured prompts and creative activities. It may offer useful support to some disabled and neurodivergent learners, depending on their individual preferences and needs.

But accessibility cannot be assumed. A tool may not work well with assistive technology, may present information in a confusing format or may misinterpret a pupil's communication style. A system trained on limited or unrepresentative data may also produce unfair or unsuitable responses.

Schools should ask:

• Has the tool been tested with pupils who have different learning and communication needs?
• Can it work with the assistive technology pupils already use?
• Can teachers adapt the output to an individual pupil's needs and learning goals?
• Are pupils and families involved in evaluating the tool?
• Is an effective non-digital alternative available where needed?
• Could the tool reinforce stereotypes or make unfair assumptions about a pupil?

AI should not be used as a shortcut for understanding a child. A pupil's needs, abilities and aspirations should not be inferred from an automated label or a single generated assessment.

Responsible education technology should widen participation and support individual progress, not create another barrier for pupils who already face disadvantages.

6. Pupil data, privacy and safeguarding

Schools hold sensitive information about children, including their contact details, learning needs, attendance, behaviour, family circumstances and, in some cases, health information. Entering identifiable pupil information into an unapproved AI service can expose that information to unnecessary risk.

Schools should follow applicable data protection law, safeguarding requirements and their own approved policies. They should assess the purpose of each tool, what information it collects, how it is stored and whether the provider uses it for other purposes.

Before adopting a product, school leaders should consider:

• Whether the school has an appropriate lawful basis for processing the information.
• Whether a data protection impact assessment is required.
• What happens to prompts, uploaded work, photographs, voice recordings and generated outputs.
• Whether the supplier retains information or uses it to train models.
• What access controls, deletion arrangements and security measures are available.
• Whether the product meets the Department for Education's relevant product safety expectations.
• How concerns, inaccurate outputs and safeguarding incidents will be handled.

Staff and pupils should not be encouraged to upload confidential information simply because an AI tool appears convenient. Schools should use approved services and explain the rules clearly.

7. Protecting learning and the teacher–pupil relationship

Education is not simply about producing a correct answer. Children need to develop knowledge, reasoning, confidence, communication and the ability to solve problems independently.

If AI completes every difficult task, pupils may lose opportunities to practise skills and discover what they can do themselves. At the same time, banning all AI use may prevent pupils from learning how to use a technology that is increasingly present in society and the workplace.

Schools need clear, age-appropriate expectations about when AI can be used, when pupils must work independently and how they should acknowledge assistance. Assessment methods may also need to focus on understanding and process, rather than relying solely on a final piece of work.

Most importantly, technology should not weaken relationships between teachers and pupils. Encouragement, professional judgement, feedback and an understanding of a child's circumstances remain essential parts of education.

8. The digital divide: who risks being left behind?

AI may create opportunities, but its benefits are unlikely to be shared equally unless schools have the necessary foundations.

Schools and pupils may differ in their access to reliable devices, high-speed internet, appropriate software, technical support and staff training. Some families may have access to paid AI services while others rely on free tools with different features or limitations.

Schools should not assume that every pupil can complete an AI-based activity at home. Nor should they make access to a particular commercial service a hidden requirement for success.

Public investment should consider infrastructure, training, accessibility and technical support alongside the cost of the software itself. It should also examine whether pupils from disadvantaged backgrounds are benefiting in practice.

9. Who receives the funding—and how should success be judged?

Public spending on AI in education can support technology developers, research, infrastructure, teacher training and school-level adoption. Partnerships between government, education providers and private companies can help develop tools for real classroom needs.

However, funding should come with clear expectations. Schools, parents and taxpayers should be able to understand what a programme was intended to achieve and whether it delivered.

Useful questions include:

• How much funding was announced, awarded and actually spent?
• Which organisations received contracts or grants, and what were they required to deliver?
• Were teachers, pupils, parents and SEND specialists involved in designing and testing the tools?
• Was the technology independently evaluated in realistic school settings?
• Did it improve learning, accessibility or teacher workload without creating unacceptable risks?
• Were benefits shared across different school settings and pupil groups?
• What happened when tools did not meet the required standards?
• Will evaluation findings and lessons from unsuccessful projects be made publicly available?

Good accountability means reporting outcomes, not just announcing spending. Where the evidence is uncertain, that uncertainty should be made clear.

10. A practical checklist for schools considering AI

1. Start with the educational need. Identify the problem before choosing a product.
2. Review the evidence. Look for independent evaluation relevant to the intended age group and use case.
3. Consult the people affected. Include teachers, support staff, pupils, families and SEND expertise.
4. Check safety and privacy. Review data handling, security, safeguarding and the relevant product safety expectations.
5. Train staff properly. Provide time to learn, practise and check outputs.
6. Set clear boundaries. Explain permitted uses, prohibited uses and how AI assistance should be acknowledged.
7. Keep human oversight. Staff must retain responsibility for professional decisions and the quality of educational resources.
8. Measure the right outcomes. Track learning, workload, accessibility, pupil experience and any unintended harms.
9. Review regularly. Reassess the tool as evidence, technology and pupil needs change.
10. Be willing to stop. Withdraw or change a tool if the benefits do not justify the risks or costs.

11. The AM-MediaSpace perspective: AI should support learning, not replace it

At AM-MediaSpace, responsible AI means treating technology as an assistant, checking important information and keeping meaningful human oversight.

In education, this means using AI to support teachers and learners while protecting children's rights, privacy, creativity and individual development.

It also means recognising that disabled and neurodivergent pupils should be included in the design and evaluation of education technology. Their needs and experiences should shape decisions, rather than being considered only after a product has been introduced.

Public funding should help schools build the knowledge, infrastructure and safeguards needed to use AI well—not create pressure to automate learning before its value has been demonstrated.

Conclusion: investment must lead to better education

AI may help schools reduce certain administrative tasks, develop learning resources and explore new forms of personalised support. Emerging evidence suggests that some tools can save teachers time, but the impact depends on how they are implemented and what happens to the time saved.

The strongest approach combines appropriate technology with well-trained staff, reliable evidence, effective safeguarding, strong privacy protections and inclusive design.

Government and technology providers should be transparent about funding, delivery and results. Schools should be supported to make informed decisions, and pupils should be able to benefit from technology without sacrificing safety, independence or the human relationships that make learning meaningful.

The goal should not be to put AI into every classroom. It should be to use the right tools where there is a clear, demonstrable benefit for teachers and learners.

Sources and further reading

• Department for Education — Generative artificial intelligence in education
• Department for Education — Using AI in education: support for school and college leaders
• Department for Education — Generative AI: product safety standards
• UK Government — AI tutoring tools could benefit up to 450,000 disadvantaged pupils (26 January 2026)
• UK Government — EdTech and AI companies invited to help build safe AI tutoring tools (16 April 2026)
• Education Endowment Foundation — Evaluation of Aila, an AI lesson-planning assistant
• Education Endowment Foundation — ChatGPT in lesson preparation: Teacher Choices trial
• Information Commissioner's Office — The Children's Code and education technologies
• Department for Education — Keeping children safe in education
• UK Government — Data and AI Ethics Framework