AI in Education: How Personalised Learning and Automated Grading Are Changing UK Schools

The school day looks different in 2026. AI-powered tools now track where each student is struggling, adjust difficulty in real time, and give instant feedback without the student waiting for a marked paper to come back. In secondary schools from Birmingham to Edinburgh, automated grading systems are cutting hours of marking every week. Some teachers love it. Others are watching carefully. Here is what the technology actually does — and what it means for UK education.

What Personalised Learning Actually Means

Traditional teaching moves at the pace of the class. Students who grasp a concept quickly sit bored while others catch up. Students who are lost fall further behind while the lesson presses on. AI changes that equation significantly.

Personalised learning systems track individual performance in real time. If a student repeatedly gets quadratic equations wrong, the system flags it, adjusts the difficulty of questions, and introduces worked examples tailored to that specific gap. If another student breezes through, they get harder material immediately. No waiting. No embarrassment in front of classmates.

When I first looked into how adaptive learning platforms work, what struck me was the granularity. A good system does not just know that a student struggles with fractions — it identifies that they struggle specifically with dividing fractions by fractions, but handle fractions by whole numbers fine. That level of detail is impossible for a teacher managing 30 pupils simultaneously.

UK Schools Leading the Way

The UK startup Century Tech, backed by Pearson, already works with over 1,500 UK schools. Their platform uses AI to create what they call a “learning map” for each student — a constantly updated picture of knowledge gaps and strengths. Teachers receive a dashboard showing exactly which pupils need intervention before they even raise a hand.

Third Space Learning, another UK company, uses AI to support one-to-one maths tutoring sessions. The AI handles scheduling, analyses session transcripts, and suggests what to cover next. The human tutor focuses on explaining and encouraging. Between 2020 and 2025, over 200,000 UK primary school pupils used their service — predominantly in schools eligible for pupil premium funding.

Khan Academy’s Khanmigo, now fully available in the UK, takes a Socratic approach. Students ask questions and the AI responds with follow-up questions rather than direct answers — designed to build understanding rather than just provide the solution. It is available around the clock, which matters for students who have questions at 10pm when their teacher is unavailable.

Automated Grading: Where It Works and Where It Falls Short

Marking is one of the most time-consuming parts of teaching. A secondary school English teacher might spend 6 to 8 hours a week marking essays alone. AI can handle a meaningful chunk of that work — with some important caveats about what it can and cannot do reliably.

For multiple-choice and short-answer questions, automated grading is well established. The system compares student responses against answer keys, applies semantic matching, and scores instantly. Teachers have used this kind of basic automation for years and few dispute its usefulness.

The harder challenge is essay marking. In 2025, several GCSE exam boards began trialling AI-assisted marking for English and history essays. The AI does not replace the human examiner — it provides a preliminary score and highlights specific phrases, which the examiner then reviews. This cuts marking time by roughly 30% per paper while keeping a human in the loop.

Early evidence suggests concerns are legitimate. Several automated marking systems have been shown to favour verbose essays with complex vocabulary over shorter, more precise answers. Marking a Year 10 essay by paragraph length rather than quality of argument is not genuine assessment — it is pattern-matching dressed up as feedback.

What Teachers Think — And What the NEU Says

The National Education Union has raised concerns about AI in schools. Their position is nuanced: they support technology that reduces teacher workload but oppose systems that replace professional judgement with algorithmic decisions. UK teachers keep asking whether these tools actually understand a student’s essay, or simply match vocabulary patterns.

Ofsted’s 2025 inspection framework acknowledges AI tools in schools but provides no specific guidance on how to evaluate them. The Department for Education published a generative AI guidance document in 2024, encouraging experimentation while stressing that teacher oversight must remain central. That tension — between “experiment boldly” and “maintain professional control” — runs through every school that has adopted these platforms.

In practice, many teachers report that the best uses are the least glamorous ones: AI setting and marking 10-question quizzes, AI generating practice exercises differentiated by ability group, AI flagging pupils who have not attempted homework. These are not exciting applications. They save real time.

AI and Special Educational Needs

One area where AI shows genuine promise is special educational needs. Text-to-speech tools, AI-powered captioning, and adaptive reading platforms all improve accessibility for pupils with dyslexia, ADHD, and other learning differences. Microsoft’s Immersive Reader and tools built into Google Classroom now incorporate AI features specifically designed for SEN pupils.

In 2025, roughly 1.6 million pupils in England had identified special educational needs, according to DfE figures. AI-powered adaptive platforms can give SEN pupils more time-on-task with material calibrated to their level — without requiring one-to-one teacher time that many schools simply cannot provide. The technology does not replace specialist SEN teachers, but it can extend their reach.

Reading tools that adapt font size, spacing, and background colour based on a pupil’s preferences — and that remember those preferences automatically — are a practical example of personalisation that helps specific groups disproportionately.

Data Privacy and GDPR

UK schools that adopt AI learning platforms must comply with UK GDPR. Every platform handling pupil data must have a Data Processing Agreement in place. Pupil data — particularly information about learning difficulties or academic performance — is classified as sensitive personal data requiring explicit handling controls.

In practice, this creates complications. Several US-based edtech platforms store data on American servers. Schools must check whether adequate data protection standards apply under UK transfer rules, and many have found they do not. The Information Commissioner’s Office has issued guidance, but compliance is uneven across local authorities.

When I looked into this, several smaller Multi-Academy Trusts were using AI tools without any formal data processing agreement in place. This is a significant legal risk — and a reputational one if a breach occurs during an inspection or a data subject access request.

The Equity Problem

Not every school can afford these tools. State schools in England operate on tight budgets. Century Tech costs money. Third Space Learning tutoring costs money. Schools in more deprived areas are least likely to have the budget for AI learning tools — and potentially most likely to benefit from them if costs could be brought down.

There is also the question of digital access at home. AI-powered homework platforms require a reliable internet connection and a device. In 2024, approximately 9% of UK pupils still lacked reliable home internet access, according to Ofcom data. An AI-powered personalised learning system is no use to a student who cannot access it after school.

The government’s EdTech Strategy (first published in 2019, updated in 2022) identified digital access as a priority but has not resolved the funding gap. The risk is that AI in education becomes one more advantage for already-advantaged schools, widening rather than closing the attainment gap.

What Universities Are Doing Differently

Universities are wrestling with AI from two directions at once. On one side, they are exploring it as a teaching and assessment tool. On the other, they are dealing with students using it to write their essays and dissertations.

Newcastle University, UCL and the University of Edinburgh have all announced AI literacy programmes as of 2025 — teaching students how to use AI tools responsibly rather than banning them outright. The consensus among most UK universities is that a blanket ban is unenforceable. Better to teach responsible use and redesign assessments accordingly.

In terms of teaching, AI tutoring tools are used in several UK universities for STEM subjects. Platforms allow students to ask questions at any hour and receive guidance that prompts them to think rather than just delivering the answer. For large lecture cohorts of 300-plus students, this kind of scalable one-to-one support is otherwise simply impossible.

What This Means for You

If you are a parent of school-age children, ask your school which AI tools are in use and how pupil data is handled. Schools should be able to answer this. If they cannot, that is itself useful information.

If you are a teacher, the tools that save the most time — automated quiz marking, adaptive practice platforms — carry the least professional risk. Use them carefully and remain the professional making judgements about student progress. Be sceptical of anything that claims to fully replace your assessment.

If you work in higher education, the AI policy question is already urgent. Institutions that leave it to individual lecturers will create inconsistency. A clear institutional framework, developed with student input, is better than a patchwork of individual responses that students cannot predict.

AI in education is not going away. The technology is improving and costs are falling year on year. The question for every UK school and university in 2026 is not whether to engage with it — it is how to do so in a way that genuinely serves students rather than just generating dashboards for senior leadership.

This article is for educational purposes only and does not constitute financial advice. Cryptocurrency investments involve significant risk. Always do your own research.