Artificial advantage?

Kevin Latham, Rebecca Montacute

Artificial intelligence (AI) has the potential to bring about transformative change in practically every sector, including in the education system. In schools, we are just starting to see its impact, with AI tools already saving teachers time in tasks like planning and marking, as well as wider staff time spent on administrative tasks. In the long term, AI could fundamentally change how education is delivered.

But the emergence of this technology also brings considerable challenges to education. While some of these have received lots of attention, such as the impact of AI on pupils’ critical thinking skills and their ability to problem solve, others are yet to be explored in great detail. This report looks to open up one area of debate regarding AI in education left relatively unexplored to date: socio-economic disadvantage.

As AI transforms the education sector, what will it do to the existing structural inequalities in the system? How will its impact be felt by different groups of students, and particularly those from lower income homes? Will the upheavals brought by AI in the education system be positive for those who are not currently served well by the status quo, or could they potentially deepen existing disadvantages?

This research explores how AI is being used in schools today, and whether inequalities in access are already starting to emerge. It looks at the barriers teachers face in making the most of the technology, and how that differs by both school type and the level of deprivation of students the school serves.

62%

The proportion of teachers who reported using AI at least once within the previous month.

55%

The proportion of schools who do not have anyone with clear responsibility for monitoring how AI is being used in their school.

Private schools are three times more likely than state schools (27% vs 9%) to have a clear school-wide staff strategy on using AI.

Key Findings

Most teachers are now using AI in schools. 62% of all teachers had used AI tools at least once within the previous month. Only 16% said they never use AI, down from 41% in April 2024.
Private school teachers are more likely (18% vs 11%) to use AI tools at least once a day. Only 8% of private school teachers do not use AI, compared to 17% in state schools.
Private school teachers (45%) are more than twice as likely to have had formal AI training than state school (21%) and 32 percentage points (77% vs 45%) more likely to have had informal training.
More than half of schools do not have anyone with clear responsibility for monitoring how AI is being used in their school – 42% of private schools compared to 57% of state schools.
Private schools are three times more likely than state schools (27% vs 9%) to have a clear school-wide staff strategy on using AI.
There are also gaps within the state sector. Teachers in schools with the most affluent intakes were more likely to report having had formal training than the least affluent (26% vs 18%).
Teachers in schools rated ‘Outstanding’ are three times more likely to have received formal training and three times more likely to report a school-wide strategy than schools rated ‘Requires Improvement/Inadequate’.
Only two fifths (42%) of teachers said they were either very or fairly confident using AI tools in their role. This rose to 51% in private schools. In state schools, 24% said they were not at all confident.

Recommendations

For government

Government should provide additional training and guidance to school leaders on the use of AI. Many teachers and senior leaders are not confident using AI tools. Existing guidance is limited, relying too heavily on judgement calls by school staff, with the potential for significant variation between settings. That the AI sector is moving apace is a challenge for everyone, but government should work with the education sector drawing on good practice and evidence to provide updated additional guidance and training for teachers and schools.
Government should tackle the digital divide for children and young people, to open up access to the benefits of AI. Many young people from lower socio-economic backgrounds still lack access to digital devices, like laptops for homework. With the growth of AI, the damage of this digital divide is only likely to grow. All students eligible for Pupil Premium should have an entitlement to a device.
The evidence base of what works on AI in education should be further developed. Often, schools are using technology without strong evidence or robust evaluation data on which approaches and tools work. Government should support research to improve the evidence base as new technologies emerge, as well as incentivising technology companies to collaborate with researchers to evaluate the educational and pedagogic impact of their products.
Government should monitor the emerging use of AI in schools, with particular attention to inequalities in access and use. This is a rapidly evolving sector. Schools who do have the capacity and ability to experiment and develop their approaches should be encouraged to do so, but with systems in place to ensure that successful new approaches can be spread to schools across the state sector. Government should regularly survey teachers and school leaders on their use of AI and monitor emerging disparities, and respond quickly before major inequalities become embedded.
Government should ensure training on digital literacy is included in initial teacher training programmes so that new generations of teachers are already equipped to teach in AI-enhanced educational contexts, while also being able to bring the latest AI-enhanced pedagogical methods into schools. They should provide ways for participants to understand the risks and see examples of where AI can support teaching or reductions in workload.
Every school should be required to have a member of staff with responsibility for AI, who should be on the senior leadership team. They should be given the time, recognition and resources to upskill as necessary, carefully formulate and implement school AI strategies and support staff with appropriate training.
The upcoming curriculum review should take account of the impact of AI on teaching and learning. This should include forward-thinking curricula that will prepare pupils for lives strongly shaped by AI both in and out of work, including benefits and risks of AI. A national AI strategy for schools should ensure pupils have equal opportunities in the workplace and beyond.
Pilots of AI tools in schools should prioritise the needs of disadvantaged pupils and narrowing gaps. This could be in the selection of tools or the location of a pilot. This report finds an emerging gap between school types in training and preparedness for AI. Schools with more disadvantaged pupils struggling to work with AI safely and effectively should receive priority support.

For schools and colleges:

Schools should designate at least one senior member of staff to have responsibility for the use of AI tools in the school. AI is complex, and the range of tools available constantly growing. Busy teachers need someone to go to for coordinated, planned guidance and support to ensure they make the most of AI’s potential.
Informed by evidence and effective pedagogy, schools should reflect on how they will monitor the use of and access to AI among students to identify the emergence of any socio-economic gaps. This should include pro-active intervention wherever gaps emerge, for example by using Pupil Premium funding to support students to access proven AI related programmes and initiatives. Financial investment in systems to help with this may be necessary.
Additional guidance and training should be provided to staff on which AI tools to use, how and with what limitations. This training should focus on the benefits that AI can bring but also clearly identify the risks of deploying it in an unregulated fashion. In MATs this guidance may be most effectively produced centrally by specialist staff.
Schools should be led by the research evidence when choosing AI-based programmes and interventions for students and staff. There is a huge number of AI-tools now being advertised to schools but with a growing number of studies being organisations like the Education Endowment Foundation, staff should be led by research evidence and privilege independently evaluated and evidenced products.
AI training should be embedded into CPD for all staff. This should minimally cover the existing Department for Education online training materials and modules, but would ideally include accompanying training focused additionally on the particular needs of a school, its staff and pupils.

Artificial advantage?

For government

For schools and colleges:

Related Research

School Funding and Pupil Premium 2025

Unequal Treatment?

Lessons Learnt?