87% of young people use AI. 30% say their school has clear rules about it
What happened
A report published on 22 September 2026 argues that England is, in its words, "trying to educate for an AI-less world that no longer exists". It was written by the historian and former head teacher Sir Anthony Seldon and Tim Bunting, a founding partner of the venture capital firm Balderton Capital, for the research company Public First. The report surveyed more than 4,000 young people aged 11 to 24 in England. The results describe a gap between what pupils are doing and what schools have set up. 87% have used an AI tool. Among secondary pupils, 63% have used AI for homework, and 77% have used it to understand something they were stuck on. Only 30% said their school had clear, explainable rules about AI. 38% said they were more skilled with AI than their teachers. Fewer than one in ten said schools or universities were preparing them very well for a world with AI. The report also cites Royal Society data indicating that only 14% of teachers feel confident in their AI literacy. Its recommendations are substantial. A rapid response unit inside the Department for Education, led by an external expert, taking on responsibilities from 1 November 2026. A national framework for adopting AI in education. Every school to have an approved AI strategy by September 2027, with staff and pupil "AI champions". Ofsted to judge schools on the quality of that strategy and to downgrade those without a robust approach. A "two-lane" model of assessment combining traditional exams with continuous, AI-supported assessment. And federated data models to track pupil progress while keeping it anonymous. Sir Anthony Seldon said AI "could offer an unparalleled category of opportunity for learning", but warned that "schools, colleges and universities have not been readied for AI's arrival".
Why this matters
The most striking pair of numbers is 87% and 30%. Nearly nine in ten young people are using these tools. Fewer than a third say their school has explained clearly what is and is not allowed. That gap is not neutral. It produces exactly the situation everybody says they want to avoid: pupils working out the rules from each other, teachers making inconsistent judgements, and honest pupils disadvantaged relative to those willing to push their luck. The 77% figure deserves more attention than it usually gets. More secondary pupils reported using AI to understand something they were stuck on than to do their homework. That is closer to what a patient tutor does than to cheating, and any policy that treats all AI use as misconduct will miss it. The teacher figure is the uncomfortable one. If only 14% of teachers feel confident in their AI literacy while 38% of pupils think they know more than their teachers, schools are being asked to supervise something many staff have not been trained on. That is a resourcing problem as much as a policy problem, and no framework fixes it without time and training built in.
What the evidence actually says
The survey is substantial. More than 4,000 young people aged 11 to 24 is a solid sample, and the findings about usage are consistent with other recent research showing AI use among young people is now normal rather than unusual. But it is worth being clear about what kind of document this is. This is an advocacy report, not a neutral evaluation. It was written by authors who believe schools are behind and who propose a specific programme of change. One author is a venture capital investor, which does not invalidate anything but is relevant context when reading recommendations about technology adoption in public services. The survey measures what young people say about their own behaviour and about their teachers. Self-reported skill comparisons in particular should be treated lightly. Thirty-eight per cent of teenagers believing they are better at AI than their teachers tells you about confidence as much as competence. The recommendations go well beyond what the survey evidence supports. Nothing in the polling demonstrates that Ofsted downgrading schools without an AI strategy would improve education. That is a policy proposal, and a contested one. Schools are already inspected against a great deal, and adding a new judgement carries real workload and risk. The "two-lane" assessment idea is likewise an idea rather than a tested method. Continuous AI-supported assessment raises significant questions about fairness, data protection and gaming that the report cannot settle. The survey finding is strong. The prescription is a proposal.
Practical advice
If you are a parent or carer, here is what is worth doing now rather than waiting for a national framework. - Ask your child's school what its rules on AI actually are. If only 30% of pupils think the rules are clear, asking is reasonable. - Ask your child how they use it. The honest answer is often revision and explanation rather than copying, and that is a much more useful conversation than an accusation. - Talk about the difference between using AI to understand something and using it to avoid understanding something. That distinction is the whole issue. - Remind them these tools get things wrong, confidently. Made-up facts, wrong dates and invented sources are common. Anything that matters should be checked. - Be careful with personal information. Do not put names, addresses, medical details or photographs of other children into these tools. - Check age requirements. Many AI services set minimum ages, and some have parental controls. - If homework has become an argument, focus on what your child can explain without the tool. That is the honest test of whether learning happened.
What to know
A report by Sir Anthony Seldon and Tim Bunting for Public First, published on 22 September 2026, surveyed more than 4,000 young people aged 11 to 24 in England. It found 87% have used an AI tool, 63% of secondary pupils have used AI for homework and 77% to understand something they were stuck on, but only 30% say their school has clear rules. Royal Society data cited in the report puts teacher confidence in AI literacy at 14%. It recommends a rapid response unit in the Department for Education, a national adoption framework, an approved AI strategy in every school by September 2027, Ofsted judgement on those strategies, and a two-lane approach to assessment. The usage findings are well evidenced. The recommendations are proposals from an advocacy report, not tested policy, and Ofsted downgrades in particular would be a significant change. For families, the useful step now is simply asking the school what its rules are, and asking your child how they actually use these tools. Sources: Schools Week, "'Rapid response unit' needed to overhaul AI in schools, says report", 22 September 2026, https://schoolsweek.co.uk/rapid-response-unit-needed-to-overhaul-ai-in-schools-says-report/; Tes, "AI-savvy students may not see school as 'legitimate use' of time", 22 September 2026, https://www.tes.com/magazine/news/general/ai-savvy-students-may-not-see-school-legitimate-use-time This article is for general information and does not replace advice from a doctor, pharmacist, teacher or other qualified professional.
Related articles
Mondays, Wednesdays and Fridays: the hub that opened at a football ground
On 12 October 2026, the Department for Work and Pensions announced that a new Youth Hub had opened in Exeter. The hub is for people aged 16 to 24. It is at St…
A 'veneer of progress'? The youth justice chief answers her critics
On 12 October 2026, the chief executive of the Youth Justice Board for England and Wales published an article defending the approach that now shapes how the sys…
Three years of missing accounts, and two trustees on the register
On 12 October 2026, the Charity Commission announced that it has opened a statutory inquiry into an education charity called Crossfields Institute. The charity…