AI safety and data
What the AI sees, and what it never touches
Written for a data protection officer as much as for a teacher. If anything here is not clear enough to put in front of your governors, tell us and we will fix the wording.
The rules our tools work to
-
No pupil data is needed
Our tools ask about a class, not about a child. There is no field anywhere that asks for a pupil’s name, and you should not type one into a free-text box either. A tool that does not need personal data is safer than one that promises to look after it.
-
The model is a witness, not an authority
Anything the model produces is a draft for you to accept, edit or throw away. It never writes into your records on its own, and it never marks anything as done.
-
References are held, not invented
A plan only cites a curriculum statement when you have picked one we actually hold for your curriculum and year group. It will not manufacture a reference to look official.
-
You can see what it was told
Every generation keeps the inputs you gave it and the version of the plan it produced, so you can see how a result was arrived at rather than taking it on trust.
Questions a data protection officer asks
Is our content used to train an AI model?
No. What you type into a tool is used to answer that request. It is not used to train a model.
Where is our data held?
Your account, your plans and your records are held in the Teach HQ platform. The privacy notice sets out who processes what, and we will provide the detail your DPIA needs on request.
Can we turn the AI tools off for our staff?
Yes. Access to each tool is controlled by your organisation, so a school can decide which tools its staff can open.
What happens if the AI gets something wrong?
You are the professional in the room. Everything a tool produces is a draft, editable in full, and nothing reaches a pupil without you deciding it should.
Do you have a DPIA we can use?
Ask us and we will send what we hold, along with a named person to answer the questions it raises.