If you're revising for GCSEs or A-Levels in 2026, chances are you've already had a go at using ChatGPT, Google Gemini, Microsoft Copilot or Claude to help you understand something. Ofcom's 2026 Children and Parents: Media Use and Attitudes report (21 May 2026) found that 56% of UK 8 to 17 year olds had used AI, rising to two-thirds of 16 to 17 year olds, and half of those users (51%) used it for schoolwork or homework.
That matters, but not always for the reasons the headlines suggest. The rules on AI in schools and exams are more nuanced than "banned" or "allowed", and the evidence on whether AI helps you learn is still thin. This piece pulls together what the primary sources say, so you can decide how much of a role AI should play in your revision without falling foul of the rules or your own memory.
The short version: studying with AI is fine. Submitting AI-generated work as your own is malpractice under JCQ rules. The interesting bit is everything in between.
What UK students are using AI for in 2026
The most reliable UK data on student AI use comes from two sources. Ofcom's 2026 report gives us the teen picture. HEPI's Student Generative AI Survey 2026 (fieldwork December 2025, published 12 March 2026, n=1,054) gives us the undergraduate one. There's no equivalent survey that covers KS4 and KS5 revision specifically, which is worth flagging.
Among UK 8 to 17 year olds, 56% had used AI. Ofcom found teenage AI users split on whether they would trust an AI-written news article as much as a human-written one. Among UK undergraduates, 95% used AI in some form during 2025/26 and 94% used generative AI for assessed work. 12% said they had included AI-generated text directly in assessed work, up from 8% the year before and 3% in 2024.
What we don't have is a clean number for how many Year 11 and Year 13 students use AI as part of their revision routine. The habit is clearly there, sitting somewhere between the Ofcom teen figure and the HEPI undergraduate one, but the specific KS4/KS5 revision benchmark hasn't been published yet.
What the rules say: DfE, JCQ and Ofqual
Three UK bodies set the framework, and they don't say quite the same thing.
The Department for Education's Generative Artificial Intelligence in Education paper (last updated 12 August 2025) sets the tone. Its position is that generative AI "cannot replace the judgement and deep subject knowledge of a human expert", and that pupil use should be "carefully considered and assessed". It also expects schools to comply with the age restrictions set by AI tools, which vary: some require users to be 13 or over with parental consent, while others, such as Claude, are 18+.
JCQ's AI Use in Assessments guidance (30 April 2025) is the one that matters most for coursework and non-exam assessment (NEA). It treats "copying or paraphrasing sections of AI-generated content so that the work submitted for assessment is no longer the student's own" as malpractice, and "failing to acknowledge use of AI tools when they have been used as a source of information" as malpractice too. Where AI is permitted, students must reference the tool, its web address and the date, as in JCQ's example: "ChatGPT 3.5 (https://openai.com/blog/chatgpt/), 25/01/2025", and keep a non-editable copy (such as a screenshot) of their prompts and the AI's responses.
Ofqual's malpractice statistics tell you how often this gets caught in practice. Ofqual recorded 100 cases of AI-related plagiarism in summer 2025, up from 85 in 2024. That was three-quarters (75%) of all student plagiarism cases but only 2% of the 5,025 student malpractice cases. Mobile phone and communication device offences remained the largest category at 44.3%. AI is a real category, but it isn't the flood the discourse implies.
Where AI genuinely helps revision
The uses students report most consistently, per HEPI 2026, are explaining concepts, summarising material and suggesting research ideas. Those are the ones with the clearest fit for GCSE and A-Level revision, provided you're doing the thinking and it's doing the scaffolding, not the other way round.
A few patterns tend to work in practice.
Explaining a concept in different ways
If your textbook explanation isn't landing, ask AI to explain a topic three different ways: at a beginner level, using an analogy, and step by step. Once one version clicks, close the tab and try to teach it back out loud. That's where the memory forms.
Turning spec bullets into practice questions
Paste a bullet from your exam board specification and ask for five practice questions in the style of that paper. Answer them without AI, then check the mark scheme. This gives you retrieval practice on the exact content examiners can test, though board-specific mark schemes are where hallucinations creep in.
Marking your own answers
Give it the mark scheme, then your answer, and ask it to mark against each criterion. The feedback can be useful for spotting missed points, but treat the grade as approximate. Board-specific mark schemes for essay questions still trip AI up.
Feedback on your writing
For English or History, drop in a paragraph you've written and ask what's weak: unsupported claims, missed analysis, structure. Do not ask AI to rewrite it for you. Once you've submitted AI-edited prose as your own, you're in JCQ malpractice territory.
Planning a schedule from your spec
Give it your exam timetable and the exam board spec, and ask it to build a topic-by-topic revision schedule. The schedule itself doesn't matter much, but the process of seeing everything laid out often surfaces topics you'd forgotten about.
Where AI trips students up
The two failure modes to watch for are hallucination and voice.
Hallucination is the AI confidently making things up. It's worst on niche spec detail: a specific required practical for AQA GCSE Physics, the exact mark scheme wording for an Edexcel A-Level History source question, a fringe case study from OCR B Geography. If the AI gets a fact wrong and you learn the wrong fact, that's worse than not learning it at all. Cross-check anything specific against your exam board's own materials, or a course built around them, like the GCSE and A-Level courses at /courses/gcse/biology-triple/aqa/higher/lessons.
Voice is the subtler failure mode: AI-drafted prose can read as generic, and teachers who know your work may notice a mismatch, but the bigger risk is that undeclared AI content in NEA is malpractice whether or not anyone spots it. This shows up in NEA and coursework more than in exam scripts. JCQ's rules on referencing exist for exactly this reason: if you've used AI, you have to say so.
Ready to boost your grades?
Join 1M+ students who have used Cognito to ace their exams.
Get started for free!
Coursework and NEA: the rules that matter most
JCQ's April 2025 guidance is quite specific about coursework and non-exam assessment. The rules most relevant to AI in coursework and NEA are:
Using AI to generate content that ends up in your final submission without declaring it. The 100 AI-related plagiarism cases Ofqual recorded in summer 2025 fall into this territory. It doesn't have to be a whole essay. Copying a paragraph, or paraphrasing a section closely enough that it's no longer your own writing, is enough.
Not keeping a copy of what the AI produced. JCQ requires that where AI is used and permitted, students keep a non-editable copy (such as a screenshot) of their prompts and the AI's responses, so a moderator can see what was AI-drafted and what was your own work. Delete the chat and you can't demonstrate the difference.
The consequences range from a written warning to disqualification from the assessment or the whole qualification, and being barred from entering exams for a set period. It doesn't happen to most students who use AI, but the framework is real.
For a broader read on what's shifting across UK and US schools this year, see AI in education 2026: What teachers, parents and students should know.
What good AI-assisted revision looks like
If you take one thing from this piece, it's that AI is at its best when it makes you do more thinking, not less. A few habits that separate students who use AI well from students who use it badly.
- Use AI to explain what your textbook or notes couldn't. Then close the tab and try to teach it back out loud without looking.
- Ask AI for practice questions from your spec. Answer them yourself. Only use AI to check your answers against the mark scheme.
- Cross-check any board-specific detail (required practicals, set texts, case studies, mark scheme wording) against your exam board's own materials or your Cognito course.
- Never submit AI-generated writing as your own. For coursework, if you use AI at all, reference it in JCQ's format and keep a non-editable copy (such as a screenshot) of your prompts and the AI's responses.
- If the revision session felt easy, you probably used AI too much. Retrieval practice should feel like effort.
The tools will keep improving. The rules will keep evolving. one of the things that won't change is that your memory only strengthens when you do the retrieval yourself. Use AI to open more doors. Walk through them under your own power. Cognito's own A-Level revision techniques guide has more on how spaced retrieval gets built into a week in practice.
