Parliament wants to ban generative AI from primary schools. This is exactly the momentum for EduGPT.

On 29 October, the House of Representatives supported, by a wide majority, a motion from Volt leader Laurens Dassen and the CDA, co-submitted by PRO, D66, Party for the Animals, DENK, 50PLUS, SP, Christian Union and JA21. The motion asks the cabinet to ban generative AI for primary school pupils. The reasoning is sharply formulated. Young pupils develop their critical thinking skills, creativity and imagination largely in primary education, and generative AI poses a risk to this.

At the same time, that very same government is working on the opposite trajectory. The new core objectives for digital literacy, which will come into force on 1 August 2027 for primary education and the lower years of secondary education, make AI literacy compulsory. Pupils must learn, already in primary school, to describe what an AI system is, how it resembles human behaviour, recognise AI applications in their own environment and use them responsibly.

A ban on generative AI and a core objective on AI literacy seem to contradict each other, but that is only an illusion. Both point in the same direction. The problem is not AI in the classroom. The problem is a child typing unfiltered into a chatbot that was never designed for education, without a pedagogical framework, without teacher supervision and without guarantees about what happens to the data. That is the risk the motion rightly warns against.

The cabinet itself also draws that conclusion, only in a different place. On Budget Day, €10.7 million was reserved for 2027 and structurally €7.4 million for a public AI facility in primary and secondary education, with Kennisnet as the lead organisation. The idea is an AI hub that acts as a switching point between applications for teachers and pupils and various underlying AI models, so that schools themselves retain control over which AI is used and under what conditions. This facility should be in place by 2028.

It is precisely that gap between now and 2028 where EduGPT already stands. Not a general chatbot where an eight-year-old can freely type, but an environment built from the ground up for education. With frameworks that match the age and curriculum, with the teacher in control rather than the pupil alone with a screen, and with the privacy and sovereignty safeguards that schools already have to take into account. That is not circumventing a ban on AI literacy, it is exactly the tool with which a school can achieve the new core objective without the risks the House is now warning about.

In short: Educational AI!

School boards currently considering how to deal with this motion and with the core objective that is set to come into force, do not have to wait until 2028. That choice can already be made responsibly now.