AI In Education Is Moving Fast, Are School Policies Keeping Up?

Artificial intelligence has infiltrated pretty much all spheres of the every day, from our professional lives to the personal too. This was greatly accelerated by the introduction of ChatGPT to everyday users, and as a result, it made its way into classrooms rather quickly. The policies governing its use, however, appear to be having a harder time keeping up – a problem being faced across industries.

And for schools, the issue of governance and regulation involves significantly higher stakes than most other sectors. The National Education Union’s 2026 State of Education survey found that 76% of teachers are now using AI tools for day-to-day work, which is up from 53% in 2025. However, 49% of teachers said their school had no AI policy whatsoever covering staff or students.

Of course, it’s news to nobody that students are hardly waiting around for schools to decide what their position on AI should be. Much like employees who jumped on the bandwagon way before their companies introduced formal policies, students started experimenting with AI not only in everyday life, but with their schoolwork too.

A recent report by historian Sir Anthony Seldon and tech entrepreneur Tim Bunting, based on a Public First survey of more than 4,000 people aged 11 to 24 in England, found that nearly 87% had experience using an AI tool. However, just 9% said they believed schools and universities were preparing them very well for a world with AI.

The report argues that England’s education system is trying to educate for an “AI-less world that no longer exists”, most likely because they’re struggling to keep up with changes rather than because they didn’t think it would become an integral part of our collective future. Indeed, this indicates that schools don’t seem to be adequately preparing young people for the reality they’re already living in, at least according to general perceptions.

The Policy Vacuum In Education

Educators are no longer just facing the challenge of whether or not AI should be allowed in schools. Now, it’s becoming more about how it how it should be used, who should be responsible for overseeing it and what safeguards need to be in place while it’s being used. Ethical AI expert Jonny Murphy-Campbell of Resolvable argues that the problem is less about AI itself and more about what happens when its use develops without proper governance.

“AI use itself isn’t a threat to children or schools, and with governance and infrastructure it could be used as a tool to vastly improve education delivery, logistics, teacher marking and feedback and ultimately how children learn and perform,” he says.

“There’s an evident significant gap between the [number] of students using AI compared to the [number] of schools without a formal policy.”

Murphy-Campbell describes this as a “policy vacuum” in education, arguing that children need both protection and the skills to use AI safely and effectively as the technology develops. Ultimately, the consequences of leaving schools to figure it out independently could be significant.

“Without a framework, you get a postcode lottery that creates unfair advantage and vision too. One school teaches pupils to question what AI tells them, while the school down the road has no idea it’s being used to write entire essays,” Murphy-Campbell says. “The risk isn’t in the fact [that] AI is dangerous, but the inconsistencies due to a lack of regulatory framework mean everyone’s making it up as they go along.”

So, if we’re going to agree that schools need to implement AI policies to help prevent these issues and inconsistencies, what should an AI policy actually cover?

That’s not a straightforward answer, but Murphy-Campbell argues that schools can at least start with some fairly straightforward principles.

A Starting Point for Schools

First, AI use should be visible. Students and staff should be clear about when and how AI has been used rather than schools simply trying to detect it after the fact. Then, there’s the question of what AI should actually be doing. The technology can support learning, but it shouldn’t replace the thinking that education is supposed to develop. Most importantly, it doesn’t need to; it has far more to offer than that.

Human oversight is particularly important when AI is used for marking or feedback. AI can assist teachers, but decisions about students should not be handed entirely to a system that lacks the context and judgement of a human educator. Thus, human-in-the-loop AI is essential when it comes to implementing AI in education.

Data is another major concern, especially since we’re considering the information of students who are mostly minors. Schools need to understand what happens to pupil information when it’s entered into AI tools, who can access it and how it is handled, and there need to be clear policies surrounding this should be done and using which platforms, models and more.

Finally, schools need to teach students how to question AI rather than simply how to use it. As Murphy-Campbell puts it, “Children must be taught – like with any other skill – that AI information is not gospel, and not always accurate.”

That may ultimately be the most important part of any school AI policy. The goal can’t just be to control which tools students can access, because realistically, that’s not necessarily even possible. Rather, it needs to be about giving them the judgement to understand when those tools are useful, when they’re unreliable and when a human should take over. Because at the end of the day, it’s not only about teaching students how to use AI in school as an isolated environment; it should be about teaching them the appropriate skills to be able to make these evaluations on other spheres of life too.

Of course, introducing and implementing AI policies in schools won’t be a quick and easy task that can be ticked off a to-do list, but putting some of the above regulations in place is a good place to start.