School leaders · 7 min
AI in UK schools: leadership, literacy and policy
Policy, AI literacy and safeguarding for SLT, governors and AI leads.
A pragmatic guide for SLT, governors and AI leads. What to put in policy, how to build staff AI literacy, and how to keep safeguarding non-negotiable.
Three things to get right first
Safeguarding — no pupil PII in any external chatbot, ever. You need a ring-fenced, protected system that can interface with SIMs, Arbor etc that never uses data for AI training and is UK law and DfE compliant
Policy — your living documents that sets expectations and direction for the AI to follow
Literacy — staff (and later pupils) using AI well, not avoiding it
What a workable UK school AI policy covers
Permitted and prohibited uses for staff
Data handling - ringfenced with your school information system
Disclosure expectations (when AI was used in pupil-facing outputs)
Pupil-facing academic integrity rules
CPD expectations and a review cycle (termly, not annually)
Building AI literacy that sticks
One-off INSET fails. AI literacy grows when staff use a tool weekly on real teaching tasks, with clear pedagogical framing — not when they're sent to a generic course.
Safeguarding non-negotiables
Never use identifiers - names, SEND details or safeguarding information in LLM prompts such as ChatGPT and Gemini etc.
Teachers.guide uses 24-hour data retention purge on solo teacher accounts; retention for school accounts is governed by the school''s DPA
Role-based access for SLT, Safeguarding Leads and class teachers
Pupil access via short code only — no pupil email
Optional school subdomain (yourschool.teachers.guide) for or inbuilt to your system for clear provenance
Related reading
For the whole-school architecture that sits behind AI adoption — ring-fenced knowledge, SLT oversight, SEND command centre, evidence-informed CPD — see School Educational Intelligence: what an EI layer is and why UK leaders need one.