Skip to main content
C&G Nest Education platform — home

Safe AI usage

Last updated 24 August 2026. This policy explains how C&G Education uses optional artificial intelligence (AI) helpers inside C&G Nest, and the rules staff must follow. It sits alongside our privacy policy and UK GDPR statement.

1. Purpose

AI in C&G Nest is a drafting aid. It exists to save staff time on session logs, planning suggestions, and plan-alignment checks. It does not replace professional, clinical, or safeguarding judgement, and it must never be the sole basis for a decision about a student.

2. Who this applies to

This policy applies to every staff member and administrator who can trigger an AI helper. Parents and carers using the family portal have no AI features and cannot see AI drafts.

3. What AI is used for

If anonymisation or the AI service fails, the platform fails open to manual drafting. Work is never blocked solely because AI is unavailable.

4. What AI is not used for

5. Anonymisation first

Before any text is sent to an AI provider, C&G Nest runs a server-side anonymisation step. Names, contact details, addresses, organisation names, and similar identifiers are replaced with neutral tokens. If anonymisation cannot be completed, the AI call does not go ahead. Raw student or staff personal data must not be pasted into an AI helper in a way that bypasses this control, and staff must not copy identifiable records into external AI tools (for example a public chatbot) for C&G Education work.

6. Human review is mandatory

Staff remain the author of every plan and log they submit. You must read, correct, and accept or discard each AI draft. You are responsible for accuracy, tone, and whether the record is appropriate to keep. Do not submit a draft you have not checked.

7. Safeguarding

If you believe a child is at risk, follow C&G Education’s safeguarding procedures immediately. Do not wait for, or rely on, an AI draft. Do not ask an AI helper whether something “is safeguarding” or to decide if a concern should be raised. Those judgements stay with trained staff and the DSL.

8. What not to put into AI helpers

9. Accuracy, bias, and limits

Generative AI can be fluent and still be wrong, incomplete, or biased. Drafts may miss context, over-state progress, or use language that does not fit the student. Treat every output as unverified. If a draft conflicts with what you observed, your observation wins.

10. Processors and location

AI helpers run on Google Cloud in the United Kingdom (London). Anonymised text may be processed by Google Vertex AI under a written processor contract. We do not use AI outputs to train a public model on C&G Education’s identifiable student records. See the UK GDPR statement for recipients and transfer safeguards.

11. Records and audit

When an AI helper is used, C&G Nest keeps an audit record of the call (time, feature, and that anonymisation ran). Submitted session logs and plans remain C&G Education’s education records, whether or not AI assisted the draft. Staff must not disable or circumvent logging.

12. Misuse

Using AI to fabricate a session that did not happen, to conceal a safeguarding concern, or to send identifiable student data to an unapproved tool is a breach of this policy and of our terms of use. We may suspend access and follow C&G Education’s disciplinary and safeguarding procedures.

How to contact us

For privacy questions, subject-access requests, or to exercise your UK GDPR rights, email office@cgeducation.co.uk.

You can also complain to the Information Commissioner’s Office (ICO), the UK supervisory authority for data protection.