What can AI do for education?
In education, AI takes over preparatory and administrative work: answering practical questions from parents and students, creating lesson materials and practice questions, giving first feedback on writing assignments and preparing timetable changes. Teaching, assessing and guiding stay with the teacher.
Besides lessons, a teacher's work consists of preparation, marking, meetings, contact with parents and administration. Part of that follows fixed steps and uses information the school already has, such as the school calendar, the book list and the course guide.
There are two ways to handle that work with AI. Teachers use an AI assistant such as ChatGPT or Claude themselves in their preparation. Or the school has an AI agent set up: a digital worker that is connected to school information and answers questions on its own. What that is, is explained in what is an AI agent.
For schools, AI Agent B.V. also has its own platform: Mentris, responsible AI for education (in Dutch). The team works in one AI environment of the school, with a knowledge base on the school's own documents. A personal data filter masks email addresses, phone numbers and Dutch citizen service numbers (BSN) before a message leaves the school. Mentris is for primary, secondary and vocational education in the Netherlands and costs from €149 per school per month (as of October 2026).
Which tasks does AI take over in education?
AI takes over four kinds of tasks: practical questions, lesson materials and practice questions, first feedback on writing and support with the timetable.
Practical questions from parents and students
"When is the parents' evening?" "How do I report my child sick?" "Which books are needed for year 3?" An AI agent answers these questions with the information the school supplies: the school calendar, the book list, the absence protocol and the school guide.
Needed: public school information. Personal data of pupils is not required for this. Human check: the agent passes questions about an individual pupil, such as grades, behaviour or care, on to the mentor or subject teacher.
Creating lesson materials and practice questions
The teacher specifies the subject matter, the level and the type of questions. With that, AI creates practice questions, a summary of a chapter, a practice test with different question types or an explanation for a wrong answer.
Needed: the subject matter and the learning objectives. Human check: the teacher checks the content for errors and for alignment with the teaching method. A language model can get a fact or a calculation wrong.
First feedback on writing assignments
AI gives a first response on the spelling, structure and argumentation of a text, using the criteria the teacher specifies. The pupil uses that response to improve the text before handing it in.
Needed: the assignment, the assessment criteria and the text, without the pupil's name. Human check: the teacher assesses the work and determines the grade. An AI system intended to evaluate learning outcomes falls under the high-risk rules of the EU AI Act, see below.
Preparing the timetable and cover
When a teacher is ill or a classroom is unavailable, an AI agent lists the possible solutions, based on availability and teaching qualification. It reports double bookings and drafts a message for the classes affected.
Needed: the timetable and teacher availability from your timetabling system. Human check: the timetabler chooses the solution and sends the message.
What can you do yourself today with ChatGPT or Claude?
With a ChatGPT or Claude subscription you can do four tasks faster yourself as a teacher today:
- Create practice questions and practice tests for your own subject matter
- Rewrite a text at a different reading level
- Draft a letter to parents or a news item based on your notes
- Work out a lesson plan for a learning objective
Kennisnet, the Dutch public organization for ICT in education, advises schools to avoid entering personal data and confidential information and to use anonymous or fictitious data where possible. So do not enter names, grades or care information of pupils. Also check which agreements your school or board has made about AI. If you have never worked with AI, start with AI for beginners.
Which rules apply to AI in education?
The GDPR and the EU AI Act apply to AI in education. This describes what those rules ask for, as of October 2026. Put your own situation to the data protection officer of your school or board.
GDPR and pupil data. Schools process data of minors. Kennisnet advises carrying out a DPIA when an AI application processes personal data and using only AI services with which agreements on the processing of personal data have been recorded. More about those agreements is in using AI in a GDPR-proof way.
High-risk uses. Annex III of the EU AI Act lists four uses in education as high-risk: AI that decides on admission or placement, AI that evaluates learning outcomes, AI that determines which level of education someone can follow and AI that monitors prohibited behaviour during tests. Since the amendment of July 2026, those rules apply from 2 December 2027 (source: European Commission). Among other things, they require human oversight.
Prohibited use. AI that infers people's emotions in an educational institution has been prohibited since 2 February 2025, with an exception for medical and safety reasons.
Transparency and AI literacy. Since 2 August 2026, people must know they are communicating with an AI system. So an AI agent that answers parent questions states that in the first message. Article 4 of the AI Act also asks organizations to take measures to develop the AI literacy of their staff. The dates and obligations are in the EU AI Act explained.
| Use | Personal data needed? | Point of attention |
|---|---|---|
| Answering practical questions | No | State that it is an AI system |
| Creating practice questions and lesson materials | No | The teacher checks the content |
| First feedback on writing | Avoidable by leaving out the name | The teacher determines the grade |
| Timetable and cover | Yes, of staff | Data processing agreement |
| Assessing tests | Yes | High-risk under the AI Act |
| Admission and placement | Yes | High-risk under the AI Act |
Where do you start per level of education?
The starting point differs per level of education, because the type of questions and the independence of pupils differ.
Primary education. The teacher handles teaching, contact with parents and reporting personally. An AI agent that informs parents about practical matters such as holidays, the school trip and PE times is a logical start here.
Secondary education. More pupils and subjects means more questions about timetables, test weeks and book lists. Start with those practical questions and then add practice questions per subject.
Vocational and higher professional education. Students have questions about internships, enrolment and regulations. An AI agent answers them with the teaching and examination regulations and the course catalogue as sources.
University. In large courses, an AI agent answers questions about the material based on what the lecturer has selected, and refers students to the right source.
Frequently asked questions
Is a teacher allowed to use AI to mark tests? AI intended to evaluate learning outcomes is listed as high-risk in the EU AI Act; those rules apply from 2 December 2027 and require human oversight, among other things. Use AI for a first response to a text and have the teacher determine the grade.
Are you allowed to enter pupil data into ChatGPT or Claude? Kennisnet advises Dutch schools to avoid entering personal data and confidential information and to use anonymous or fictitious data where possible. Use only AI services with which your school has recorded agreements on the processing of personal data.
Does AI in education fall under the EU AI Act? Yes. AI for admission, evaluation of learning outcomes, determining the level of education and monitoring during tests is high-risk, with rules from 2 December 2027. Emotion recognition in educational institutions has been prohibited since 2 February 2025.
Does AI replace the teacher? No. AI produces drafts for lesson materials, practice questions and messages, and answers practical questions. The teacher does the teaching, guiding and assessing.
What does AI cost for a school or teacher? A 1-on-1 AI training costs from €449 excl. VAT. The implementation of an AI agent starts from €1,495 one-time, on invoice (as of October 2026). Read more about what an AI agent costs.
Using AI in your organization?
Curious what a digital worker can concretely do for education? Check our AI agent for education or read what an AI agent costs. Rather upskill your team first? Our 1-on-1 AI training covers that. Or book a free intro call.
