Since the advent of generative artificial intelligence (AI), it feels like we are constantly being bombarded with new ways that the technology is going to improve all aspects of our lives. Education is no different. Indeed, the government is currently developing a £4 million AI tools programme that it promises will ease teacher workloads and improve the quality of their lessons. Unfortunately, whilst the government is clearly keen to seize the opportunities of AI when it comes to benefiting teachers, it remains worryingly quiet on how students are using AI in their own time, and has not put in place measures to train pupils to use AI responsibly.

The government’s ambition to help teachers is understandable. According to a National Education Union survey, 62% of teachers found it commonplace to work in the evenings, and 36% cancel social plans to get on top of their workload. As such, investment in technological support for teachers is both necessary and timely. The £3 million investment into a Content Store, designed to train new generative AI to be rolled out in schools, would help ease this strain, as developers estimate new AI tools can reduce time spent on pupil assessment by 50%. If implemented effectively, this is clearly a sensible move.

However, while the government is developing a training package and guidance on the safe use of AI for educators, no equivalent advice is being issued to pupils to ensure they are equipped to use AI responsibly.

According to a National Literacy Trust survey of 51,000 young people, the percentage of people under 18 who said they used generative AI increased from 37% in 2023 to 77% in 2024. While AI usage in itself is not an issue, problems arise when we realise that over 20% of participants admitted to not verifying what generative AI told them and copying the output entirely.

This misuse of AI is of growing concern to both educators and parents, as fears mount that it could harm independent learning. A study conducted by the Alan Turing Institute found that 76% of parents were worried that AI was negatively impacting their child’s critical thinking skills, and 52% of teachers are concerned that work being submitted by school pupils is AI-generated. This reflects the growing concern that AI erodes the long-term efforts of schools to harness independent thinking skills.

There is little hard evidence available to gauge AI misuse at secondary school level. However, we can get a good idea from the more concrete evidence visible in our universities. A survey of 80 UK universities found that 82% have investigated their students for using AI to cheat, with Birmingham City University reporting 402 instances of cheating between 2022 and 2024.

The potential dangers of the government’s oversight in AI misuse is further evidenced by a recent paper from the University of Edinburgh, which found that pupils were regularly using AI to solve their maths homework, despite the same students exhibiting strong opinions against using generative AI in tests.

Currently, the only formal policy guidance on this issue comes from the Department for Education’s (DfE) policy paper, ‘Generative AI in education’. This briefly references the use of AI in assessments and offers guidance from the Joint Council for Qualifications, but adds little else. The Joint Council’s advice is equally lacklustre, advocating that the teachers somehow police their students by allocating time for tasks to be completed in class under supervision, or banning AI tools on school devices. As such, schools are left almost complete autonomy to decide whether students can use AI, how they do so, and what safeguards are put in place.

This sort of laissez-faire approach may be suitable to manage phone usage in schools – the government currently leaves schools to establish their own policy in this area – but a completely hands-off approach is risky when it comes to AI.

Given that misuse of AI in exams, homework and essay writing is difficult to police, it should be a DfE priority that students are at the very least taught how to use AI in a way that enhances their education without fostering dependence on it.

Back in 2023 the Chair of the Russell Group of the UK’s leading universities stated that students must be trained to use models like Chat GPT without cheating. With children as young as seven using these tools in their formative years of schooling – 59% of 7-17 year-olds have used generative AI tools in the last year –  the government simply must do more to ensure that Britain’s youth becomes AI literate.

A first step would be to adapt the government’s existing training package, currently aimed at teachers, and extend it to all secondary school pupils, whilst ensuring the content is age appropriate.

However, more must be done. Whether this involves regular mandatory AI lessons covering its uses, foundations, and risks; or new government investment into research, it is vital that the government gains a better understanding of pupils’ misuse of this technology and is better equipped to combat such behaviours.

An important step in the right direction would be to update the 2011 Education Act to include the use of AI, with a deliberate recognition that the government must protect its youth by ensuring they can use this technology responsibly.

At a time when pressures on teachers are at an all-time high, the government is right to utilise technology to help teachers and improve their workload. But this enthusiasm must not come at the expense of oversight elsewhere. After all, pupils’ AI usage outside the classroom is just as – if not more – important than their usage in it.

George Somper is a reporter and a Member of Bright Blue. The views expressed in this article do not necessarily reflect the views of Bright Blue.