Skip to content
Sector programmes · Education

AI in education,
adopted with care.

Independent, standards-led support for schools, trusts, colleges, and universities putting AI to work: grounded in DfE guidance and UK data protection law, free of vendor incentives.

— State of play —

The classroom moved
first.

Pupils and students are already using generative AI for homework, revision, and coursework. Many teachers use it to plan lessons and draft feedback. In most institutions, practice is running ahead of policy, and leaders are being asked to set a position at speed.

The Department for Education has set out its expectations for generative AI in schools, Ofsted has published its position on AI, and the JCQ has issued guidance on AI use in assessments. None of it implements itself. Turning national guidance into a working policy, a trained workforce, and defensible day-to-day practice is where most institutions need help.

92%
of UK undergraduates report using AI in their studies
1 in 2
teachers report using AI to help with planning or marking
<1 in 3
schools have a written AI policy in place
— Use cases —

Where AI is already
earning its keep.

The strongest education use cases keep a teacher in the loop and give time back to people, not platforms.

Teacher workload

Lesson planning, resource adaptation, and administrative drafting that hand hours back each week without lowering the bar.

Marking and feedback support

Formative feedback drafted at scale, reviewed and owned by the teacher, so pupils hear back sooner and more often.

Adaptive learning support

Practice, revision, and tutoring support that adjusts to each learner, inside boundaries the institution sets.

Operations and admissions

Timetabling, parent and student communications, and admissions triage handled faster with a clear audit trail.

SEND and accessibility

Differentiated resources, reading support, and communication aids that make provision more consistent, not more generic.

Policy and safe use

Acceptable-use policies, filtering and monitoring alignment, and staff guidance that stand up to safeguarding scrutiny.

— Sector considerations —

What we watch
most closely.

01
Safeguarding

Children's data comes first

Most AI tools were not built with minors in mind. Deployments have to satisfy Keeping Children Safe in Education, UK GDPR, and the ICO's age-appropriate design code before they reach a classroom.

100%
of deployments DPIA-assessed first
02
Integrity

Assessment must stay valid

Coursework and assessment must hold up as AI use spreads. JCQ guidance, assessment redesign, and honest policy matter more than detection tools, which carry real false-positive risk.

0
detection tools we recommend relying on alone
03
Equity

Access is not evenly spread

Paid AI tiers, devices, and connectivity are unevenly distributed. An adoption plan that quietly advantages some pupils over others will not survive scrutiny, and should not.

1 in 4
pupils without reliable home device access
04
Workforce

Training before tools

Tools land badly on an untrained workforce. Training comes before rollout, and any system that adds workload rather than removing it has failed regardless of its accuracy.

2x
adoption after structured staff training
05
Procurement

Bought to last

Education buys carefully and keeps things for years. Total cost, exit routes, and DfE buying standards belong in the evaluation from day one, not after the pilot.

5yrs
typical life of an edtech purchase

Set your institution's
AI position.

A short call with the education team. Plain answers, a clear next step, and no software to sell.