Government push

In early 2025, the UK’s education secretary, Bridget Phillipson, told The Bett Show the current government plans to use AI to save teachers time, ensure children get the best education possible, and grow the connection between students and teachers.

But research conducted by the Department for Education to gauge teachers’ attitudes to the technology found many are wary. Half of teachers are already using generative artificial intelligence (GenAI), according to the research, but 64% of the remaining half aren’t sure how to use it in their roles, and 35% are concerned about the many risks it can pose.

Regardless of teacher attitudes, the government is leaning heavily into using AI to make teachers’ lives easier, making plans to invest £4m into developing AI tools “for different ages and subjects to help manage the burden on teachers for marking and assessment”, among many other projects and investments.

The Department for Education (DfE), which also commissioned Ofsted’s research into the matter, has stated: “If used safely, effectively and with the right infrastructure in place, AI can ensure that every child and young person, regardless of their background, is able to achieve at school or college and develop the knowledge and skills they need for life.”

Use cases and cautions

Early in 2025, the government launched its AI opportunities action plan, which includes how the Department for Science, Innovation and Technology (DSIT) aims to use AI to improve the delivery of education in the UK, with DSIT flagging potential uses such as lesson planning and making admin easier.

In some cases, this is exactly what schools and colleges were using it for, according to Ofsted’s research – many were automating common teaching tasks such as lesson planning, marking and creating classroom resources to make time for other tasks; others were using AI in lessons and letting children interact with it.

Other schools had already started developing their own AI chatbots, and though no solid plans were yet in place, there were hopes of integrating the technology into the curriculum in the future.

But implementing AI has required careful consideration, with the report highlighting: “AI requires skills and knowledge across more than one department.”

Each school and college Ofsted spoke to were in different stages of AI adoption, as well as teachers and students having varying levels of understanding of how best to use the technology.

Pace of adoption also varied, though most schools seemed to be taking an incremental approach to adoption, changing bit by bit as teachers and students experiment and accept new ways of working using AI technology. The report claimed there didn’t seem to be a “prescriptive” approach about what tools could be used.

An “AI champion” existed in most cases, namely someone responsible for implementing and getting others on board with adoption – usually someone who has prior knowledge of the technology in some capacity.

A college principal of one the education providers Ofsted spoke to said: “I think anybody who’s telling you they’ve got a strategy is lying to you because the truth of the matter is AI is moving so quickly that any plan wouldn’t survive first contact with the enemy. So, I think a strategy is overbaking it. Our approach is to be pragmatic: what works for the problems we’ve got and what might be interesting to play with for problems that might arise.”

When children are involved, safeguarding should be at the forefront of any plans to implement new technologies, which is one of the reasons those running pilots and introducing AI are being so cautious.

Those Ofsted spoke to already displayed knowledge about the risks of using the technology, such as “bias, personal data, misinformation and safety”, and many had already developed or were adding to AI policies and best practices.

The report said: “A further concern is the risk of AI perpetuating or even amplifying existing biases. AI systems rely on algorithms trained on historical data, which may reflect stereotypical or outdated attitudes…

“However, some of the specific aspects of AI, such as its ability to predict and hallucinate, and the safeguarding issues it raises, create an urgent need to assess whether intended benefits outweigh any potential risks.”

There have been other less commonly mentioned concerns for some schools, for example, where AI is being used for student brainstorming or individualised marking, there is the possibility of narrowing what is counted as correct, taking away some of the “nuance and creativity” from how students can answer questions and tackle problems.

“Deskill[ing]” teachers and making it harder for children to learn certain skills because of a reliance on AI was also mentioned as something education providers are worried about.

Getting it right

Ultimately, AI adoption will be an ongoing process for education providers, and it’s important senior leaders are on board, with someone in charge of introducing and monitoring the technology’s impact on teaching and education delivery.

The most vital piece of the puzzle, according to Ofsted, is ensuring teachers are guided and supported rather than put under pressure, as well as guaranteeing transparency surrounding anything AI is used for in schools.

“There is a lack of evidence about the impact of AI on educational outcomes or a clear understanding of what type of outcome to consider as evidence of successful AI adoption,” the report said. “Not knowing what to measure and/or what evidence to collect makes it hard to identify any direct impact of AI on outcomes.

“Our study also indicates that these journeys are far from complete,” it continued. “The leaders we spoke to are aware that developing an overarching strategy for AI and providing effective means for evaluating the impact of AI are still works in progress. The findings show how leaders have built and developed their use of AI. However, they also highlight gaps in knowledge that may act as barriers to an effective, safe or responsible use of AI.”