Students and AI: 40+ UK Statistics on How Young People Use AI to Study (2026)

94% of UK undergraduates now use generative AI to help with assessed work. Two years earlier it was just over half. Younger pupils are not far behind: among 8 to 17-year-olds who use AI, half (51%) use it for schoolwork or homework, up from 34% in 2023.

How they use it matters as much as whether they do. One in four teenagers who use AI (25.1%) say they “just copy” what it gives them for homework, and the share is rising. At university, by contrast, the most common use is asking AI to explain a concept.

This report brings together three large UK surveys: Ofcom on children aged 8 to 17, the National Literacy Trust on more than 60,000 teenagers aged 13 to 18, and HEPI on full-time undergraduates. Each figure links to its source.

Top Picks

Use

• Children aged 8 to 17 who have used AI: 56%, rising to 66% of 16 and 17-year-olds
• Teenagers aged 13 to 18 using generative AI weekly or more: 45.6%, up from 31.1% a year earlier
• Undergraduates using AI for assessed work: 94%, up from 89% in 2025

Habits

• Teenage AI users who use it for help with homework: 60.9%
• Teenage AI users who just copy its answers for homework: 25.1%
• Undergraduates who use AI to explain concepts: 61%
• Undergraduates who have pasted AI text straight into assessed work: 12%, up from 3% in 2024

Divides and doubts

• Children who have used AI, better-off vs less well-off households: 67% against 46%
• Undergraduates who say AI skills are essential: 68%; who feel staff are helping them build those skills: 48%

How Many Students Use AI?

Use climbs steeply with age. Ofcom found that just over a third of 8 and 9-year-olds have used AI, against two-thirds of 16 and 17-year-olds. Among undergraduates, almost everyone has.

Awareness is close to universal. 86% of 8 to 17-year-olds have heard of AI tools such as ChatGPT, Copilot or Gemini, and 96.9% of 13 to 18-year-olds have heard of generative AI.

Is Student Use of AI Still Growing?

At university, yes. The share of undergraduates using AI in at least one way went from 66% in 2024 to 92% in 2025 and 95% in 2026:

Among children, Ofcom's figure has risen from 46% to 50% to 56% over three years, although Ofcom says the trend is only indicative because its questions changed.

Teenagers are the exception. The National Literacy Trust found fewer 13 to 18-year-olds using generative AI in 2025 than in 2024 (66.5% against 77.1%), driven by 13 to 16-year-olds. But those who did use it used it more often: weekly use rose from 31.1% to 45.6%.

Fewer teenagers are trying AI, but the ones who use it are using it more.

What Do School Pupils Use AI For?

Homework is near the top for every age. Among 13 to 18-year-olds who use generative AI at least monthly, these were the main reasons:

Girls were more likely than boys to use AI for homework (66.4% against 56.2%), and 16 to 18-year-olds more likely than younger teenagers (70.4% against 59.8%). Using AI “to have a chat” halved in a year, from 44.4% to 22.4%.

Ofcom's survey of AI users aged 8 to 17 tells a similar story:

• Schoolwork or homework: 51%, up from 45% in 2024 and 34% in 2023
• Learning more generally: 43%
• Just for fun: 43%
• Finding information such as news or weather: 28%
• Making images or songs: 20%
• As someone to talk to, or “as a friend”: 11%

Do Students Copy AI Answers?

Some do, and more than before. A quarter of teenage AI users (25.1%) admitted to “just copying” AI outputs for homework in 2025, up from 20.9% in 2024. Asked about writing specifically, about 1 in 3 said they had used AI simply to produce answers to homework questions.

Many are more careful than that:

National Literacy Trust, 2025. The survey reports the two age bands as “13 to 16” and “16 to 18”.

Older teenagers were far less likely to copy: 10.0% of 16 to 18-year-olds, against 26.8% of 13 to 16-year-olds. And most still value the skill AI could replace: 65.5% agreed that “even when we have AI, it is important to learn how to write”, rising to 80.9% of 16 to 18-year-olds.

Half (50.8%) said AI helped them learn or understand things when parents and teachers weren't around. Only 15.5% said it made them feel more independent when reading or writing.

How Do University Students Use AI?

Mostly to understand, not to write. When preparing assessed work, undergraduates had used generative AI for:

Pasting AI-generated text straight into assessed work is still a minority habit, but it has grown each year: 3% in 2024, 8% in 2025 and 12% in 2026. Using AI to generate text in general fell, from 64% to 56%, which HEPI suggests may reflect a move to more specialised tools.

Universities have responded. 65% of students said assessment had changed significantly because of AI, and several described worrying about being falsely accused of cheating.

Is AI Helping or Harming Students?

Undergraduates are split, though more say it has helped:

• AI has made their student experience better: 49%
• Made it worse: 16%
• No significant impact: 35%

AI has also become a source of company for some. 15% of undergraduates used AI for friendship, company, advice or to tackle loneliness, and 8% used counselling or therapy provided solely by AI. Its effect on loneliness cuts both ways: 21% said it made them feel less lonely and 20% more lonely, while 59% noticed no effect.

Independent study time has fallen at the same time. Full-time undergraduates averaged 11.1 hours a week in 2026, the lowest on the Student Academic Experience Survey's record, and its authors suggest AI may be speeding up parts of independent study.

Do Young People Trust What AI Tells Them?

Too much, in some cases. Nearly half (46%) of 8 to 17-year-olds who have read AI-generated summaries in search results believe those summaries are always accurate. Among teenage AI users, 40% would trust an AI-written news article as much as or more than one written by a person, while 43% would trust it less.

Is There an AI Divide Between Students?

Among younger children, yes. 67% of 8 to 17-year-olds in better-off (ABC1) households have used AI, against 46% in less well-off (C2DE) households. HEPI also found undergraduates from C2DE households less likely to use some AI tools, even though free versions exist.

The National Literacy Trust found no such gap among teenagers when measured by free school meals: 66.7% of those receiving them had used generative AI, against 67.0% of those who did not. The trust notes that household social grade and free school meals measure different things, which may explain the difference.

Are Students Being Taught to Use AI Well?

Students think AI skills matter, but many say they are not being taught them:

HEPI recommends that universities give all new students structured AI training, with a focus on using AI “to enhance learning, rather than outsourcing it”.

Conclusion

AI is now part of studying at every age in the UK, from half of 8 to 17-year-olds to almost every undergraduate. The bigger question is how it is used. Older students mostly ask it to explain things; a growing minority of younger teenagers copy its answers. Students also say they want help using it well, and fewer than half feel they are getting it.

AI works best for learning when the student still does the thinking: asking it to explain a topic, then organising and recalling the ideas themselves. Our mind map and study statistics report covers the evidence on which study methods work.

How We Put This Together

Every figure was checked against the original report. The three surveys cover different ages and ask different questions: Ofcom asks whether children have used AI tools such as ChatGPT, Copilot or Gemini; the National Literacy Trust asks about generative AI, and its reasons for use cover those who use it at least monthly; HEPI asks full-time undergraduates about AI generally and generative AI for assessed work. We show them side by side rather than combining them. Ofcom describes its year-on-year trends as indicative because its methods changed.

Sources

• Ofcom (2026). Children and Parents: Media Use and Attitudes Report. Published 21 May 2026.
• Picton, I., Clark, C., & Bonafede, F. (2025). Young people's use of generative AI to support literacy in 2025. National Literacy Trust.
• Stephenson, R., & Armstrong, C. (2026). Student Generative AI Survey 2026 (HEPI Report 199). HEPI and Kortext.
• Neves, J., Stephenson, R., & Armstrong, C. (2026). Student Academic Experience Survey 2026. HEPI and Advance HE.

About the author

Kashish Mustafa, Co-founder, Mindmapi

Kashish was predicted 3s and 4s in her GCSEs. Studying on her own, she mind mapped everything, maths included, and came out with 7s and 8s. She went on to graduate with a 2:1 in Business Management and now works in marketing. She co-founded Mindmapi to give every student the study method that worked for her.

Kashish on LinkedIn