A history teacher at a secondary school pastes twenty test questions into a free tool on a Sunday evening and gets them back at four levels. On Monday a colleague in the staff room shares an extension that summarises lesson videos. Neither of them reported it, and neither broke a rule: nothing has been decided on this yet.
That is shadow AI in education: AI that staff have started using on their own, without a decision from the school. It comes from three directions, all of them out of sight of the ICT team: free tools on personal accounts, AI features that appear inside learning tools already in use and browser extensions. What the school needs is not a ban but a view of which tools are being opened, a decision on each one and that decision in front of staff when they open the tool. None of this requires tracking a single teacher.
Where does shadow AI in a school come from?
Our article on what shadow AI is gives the general definition: self-serve, no decision, actually in use. In schools it takes three recognisable forms.
| Where it comes from | Example | Why it stays out of sight |
|---|---|---|
| Free tool on a personal account | The free version of ChatGPT, MagicSchool, Curipod, a translation service | No procurement, no school sign-in, often used at home on a personal device |
| AI feature in a learning tool already in use | A pupil tracking system or digital course materials that add a summarising or marking feature | Same domain, same data processing agreement, new processing that nobody has reassessed |
| Browser extension | An extension that summarises videos, rewrites emails or checks text | Installed in two seconds, with read access to every page and no admin involved |
The first form is the best known and the easiest to spot. The second is the quietest: the tool was already in the register as allowed, and the new feature came in with an update. Our article on why tool catalogues age explains why that alone is reason to reassess. The third is overlooked almost every time, even though an extension with read access is an AI tool too, and one with a wider reach than the chatbot itself.
Why is it growing in schools in particular?
For the same reason as everywhere else: it makes the work lighter. Feedback on thirty essays, a test at three levels, a letter to parents without jargon. There is usually an approved tool, but on a Sunday evening it does not feel up to the job, or the teacher simply knows it less well than the one already open at home.
Three things set a school apart from an office. Staff do a lot of their work from home and on personal devices, so the school network sees little of it. The data involved belongs to children, and a pastoral care note or a development plan is the most sensitive of all. And schools work with data processing agreements under the education sector's privacy covenant, whereas a free account runs under consumer terms without one. In education, that is the first question for every tool that surfaces, ahead of what it can do. The Dutch Data Protection Authority has reported receiving several data breach notifications in which employees had put personal data into an AI chatbot, and it advises organisations to set clear rules on whether this is allowed and which data may go in.
A ban does not solve this. It does not remove the tool, only the tab you could still see; the use moves to phones and personal accounts. You are left with exactly the kind of account you did not want.
Why does a quarterly survey fall behind?
In October, a school board with twelve schools asks its ICT coordinators which AI tools are being used at their school. The list comes to seventeen tools, six of them without a decision. The board assesses them, records the statuses and circulates the overview in December.
By January there are three new tools, one of which turned up at four schools at once after a webinar. Over the same period, a learning tool listed in the register as allowed has gained a feature that comments on essays. And the extensions were never on the October list in the first place, because nobody recognised them as AI tools.
The survey did not fail. A list that people fill in from memory is a snapshot, and use keeps moving. Our guide on how to discover shadow AI sets out the available methods and what each of them misses.
How do you keep track without tracking teachers?
By limiting the measurement to the question the school actually has: which AI tools are being opened, and which of them have no decision yet? To answer that, the domain of the service is enough: that an AI tool was opened at the board's schools this month, which one and how often. Who opened it adds nothing, and the moment you record it anyway, a count of services becomes a record of people. That difference, and what it means under the GDPR and for staff representation, is covered in our article on AI monitoring versus employee monitoring.
That is why BeeSensible records only the registrable domain, at most once a day per device, with no user ID, no path and no page content. A new tool lands in the Inbox with the status "No decision yet"; the board's administrator assesses it and records the decision. This is what that register looks like:
