artificial intelligence is already in use across schools in the UK, yet only 2% of those surveyed have developed a full strategy for how teaching, learning and school management should be supported, a UK study found. Accenture and teach first.
The study found that 12% of schools surveyed have implemented AI policies and 20% offer AI-focused staff training. Almost two-thirds of school leaders, 63%, identified lack of staff confidence or skills as a barrier to recruitment.
The study combines a survey of nearly 200 school leaders, 30 semi-structured interviews, and six interviews with AI and education experts. This report focuses primarily on secondary schools in England.
School leaders surveyed said teachers are already using AI for activities such as lesson planning, quiz writing, and exam question writing. However, the report found that this use often develops informally, rather than through a consistent approach across schools.
Accenture and Teach First recommend that school leaders use AI themselves, setting clear objectives and boundaries, starting with low-risk applications, and providing opportunities for staff to test tools and share what they learn.
Only 2% of schools surveyed have a complete AI strategy
The report distinguishes between policies that establish rules for the use of AI and broader strategies that cover what schools want to achieve with technology, what limits it should have, and how to develop staff capacity.
12% of schools surveyed reported having an AI policy, but only 2% said they had a fully developed strategy. Accenture and Teach First explain that the remaining work is fragmented, with schools often making local decisions without a shared model or system for exchanging evidence.
Among school leaders surveyed, 60% said they use AI at least weekly. This includes 16% who use it daily and 44% who use it at least once a week. A further 14% said they do not use AI.
Most leaders were positive about the possibility. While the report found that 79% agree that AI has the potential to improve teaching and learning outcomes, interviews suggested that many were clearer about the potential risks than the specific educational uses.
Concerns include plagiarism, protection, inaccurate or biased information, data privacy, and the potential for faculty and students to become overly reliant on AI-generated work.
The most frequently reported barrier was staff competence. Some 63% of leaders cited a lack of confidence or skills among staff, 51% cited data privacy concerns, and 41% pointed to a lack of understanding of how AI can support education.
Cost, cited by 16%, is a less common concern than trust, privacy, and limited knowledge.
Matt Prebble, Head of Accenture in the UK and Ireland, said: “Artificial intelligence is no longer a future consideration for schools; it is already integrated into everyday teaching and learning. This research shows that the adoption of artificial intelligence is outpacing the strategy, training and support needed to sustain it.”
He added: “Many school leaders are responding to this change without clear guidance or confidence to implement technology effectively.”
Training and leadership shape recruitment
Currently, only one in five schools surveyed offers AI-focused training. Three in ten leaders estimated that less than 20% of teachers in their schools are confident in using AI. These numbers reflect school leaders’ ratings rather than responses collected directly from teachers.
The report found that adoption was more consistent in schools where senior leaders used AI themselves and demonstrated how it could be applied. Schools with more skeptical or less proactive leadership are said to be slower to adopt and less aligned.
This relationship does not indicate that demonstrating leadership alone causes stronger recruitment. Schools with more established digital capabilities, staff expertise, or access to technology may also be in a better position to implement AI.
The report found that there are regional differences in how instructors themselves use the system. In London, 29% of school leaders surveyed said they use AI every day, compared to 12% in other parts of England.
This comparison measures everyday AI use among leaders rather than examining classroom adoption, student access, and educational outcomes. The study does not establish whether London schools are using AI more effectively or whether students are benefiting more.
Accenture and Teach First warn that unequal access to talent and technology can widen existing disparities between schools and regions.
James Toop, chief executive of Teach First, said: “Artificial intelligence has the potential to support education and reduce workload pressures, but there is a real risk that changes in access to technology and training could deepen the inequalities that exist, particularly between schools and students outside of London.”
He added: “Too many teachers and school leaders are being forced to navigate these changes without the guidance and support they need.”
Report requires controlled experiment
School leaders reported that teachers are leveraging AI to reduce preparation time through activities such as drafting lesson materials, creating questions, and creating administrative content.
The study did not independently measure how much time teachers saved, whether the AI-generated resources were accurate, or whether the use of AI improved student achievement. The findings regarding benefits are primarily based on leadership descriptions and interviews.
Accenture and Teach First recommend five steps for schools.
The first is for leaders to engage directly with AI and demonstrate responsible use. We also recommend that schools define their objectives and boundaries before expanding access, including which activities are appropriate for AI and which activities require human judgment to remain central.
The report recommends starting with low-risk applications where value is easier to identify, while building toward broader plans. It also requires schools to allow controlled experimentation within clear professional and safety safeguards.
A final recommendation is to develop competency through continuous shared learning rather than relying solely on isolated training sessions. This could include staff comparing tools, discussing mistakes and sharing examples of effective practice.
The survey represented around 6% of the secondary education sector in England and was supplemented by interviews aimed at providing additional background. It does not claim to be a complete audit of the use of AI in all schools, nor does it include direct measurements of student outcomes.
