It’s ten past one. The toddler room has finally gone quiet, the lunch trolley has rattled back to the kitchen, and someone on your team is using naptime to catch up on paperwork. Increasingly, they’ve also got a chatbot open in another tab. It drafts the parent newsletter in thirty seconds. It smooths out a clumsy paragraph in a report. It suggests six things to do with a water tray in a heatwave.
None of that is a scandal. Used well, these tools hand practitioners back hours that end up where they belong, with the children. The trouble starts with what gets typed into the box.
How AI arrives without anyone deciding
We hear the same story from managers again and again: nobody chose to adopt AI, it just turned up. One practitioner used it to word a tricky message to a parent. Someone else discovered it could summarise a long policy document. By half term it was quietly part of how the setting runs.
That gradual arrival is exactly the problem. A tool nobody decided to adopt is a tool with no rules attached. And in early years the rules matter more than in most workplaces, because our paperwork is dense with children’s personal information.
The line that must never be crossed
Here’s the plain version. When you paste text into most AI tools, you are sending it to a company’s servers. You don’t control where it goes, how long it’s kept, or what it might be used to train. So a child’s name, date of birth, medical details, safeguarding notes or photographs must never be entered into an AI tool. Not to save time on a learning journal entry. Not “just this once” on a Friday afternoon.
UK GDPR applies to a chatbot exactly as it applies to a filing cabinet. An observation about a named two-year-old is that child’s personal data, and typing it into a free online tool is a disclosure you can’t take back.
The reassuring part is that nearly everything AI is genuinely good at works fine without personal data. Draft the newsletter with no names in it. Ask for activity ideas for “a mixed group of two-year-olds” rather than for a specific, named child. Strip the details out first, then let the tool do its party trick.
Confident, fluent and sometimes wrong
AI writes with total confidence, which is rather the problem. It will produce a beautifully fluent paragraph about ratios or allergen rules that is simply incorrect, and it won’t blush. Anything factual needs a human check from someone who knows the answer, especially anything a parent will read or an inspector might see.
There’s a bias question too. These tools learned to write from the internet, and the internet has opinions. Left unchecked, AI-drafted observations can drift towards the generic and flatten the real variety of the children in front of you. The words in a child’s record should come from someone who knows that child, even if a tool helped with the typing.
And there are safeguarding risks that every practitioner should understand at least at awareness level, from misleading generated images to over-trusting a tool’s advice on a sensitive situation. Our online course Safe Use of AI in the Early Years Setting covers those risks practically, alongside the legal duties and the everyday habits that keep AI use honest.
Write the rules down before September
Every setting needs a short, workable AI policy: which tools are allowed, what may never be entered into them, who checks outputs, and how you’ll be open with parents about what’s AI-assisted. Written now, in the calmer weeks of July, it becomes part of induction for your September starters rather than a scramble mid-autumn.
Keep it to a page. A policy staff can remember beats a policy that impresses a filing cabinet.
Clear AI rules for your whole team, in one short course.
Safe Use of AI in the Early Years Setting covers your UK GDPR duties, the golden rules for children’s data and how to write a policy that works, with an NFAQ-accredited certificate on completion.
AI isn’t going back in its box, and on a busy week we wouldn’t want it to. The settings that get this right won’t be the ones that ban it. They’ll be the ones where every adult knows exactly where the line sits.