Kennisnet provides primary and secondary education with guidance through the AI Literacy Framework
On 9 September 2026, Kennisnet published the AI Literacy Framework for primary and secondary education, together with a toolkit to help teams get started. We read it with great interest, as the framework closely aligns with how we support schools with EduGPT. In this post, we summarise the key points and explain where the connection lies.
Why this framework exists now
The European AI Regulation requires organisations, including schools, to ensure that staff are sufficiently AI literate. This obligation, set out in Article 4, has been in force since 2 February 2025. Since then, many school boards have wondered what AI literacy actually entails and how it can be demonstrated.
Kennisnet’s framework now provides an answer for primary and secondary education. It is an adaptation of the AI-GO! framework from the Npuls Growth Fund programme, which was developed for secondary vocational, higher professional, and academic education. The authors of the article are Astrid Buijs and Michel Cents.
Four competencies, three levels
The framework describes four competencies. Proficient is about understanding how AI works and what it can do. Critical is about assessing AI and its impact on people and society. Creative is about thinking, making, and collaborating with AI.
The fourth competency, Professional, will be elaborated on later. This will be tailored to each role, for example for teachers, school leaders, and support staff. For teachers, it will be based on their pedagogical and didactic knowledge and skills.
In addition, Kennisnet distinguishes three levels.
Basic level, for every staff member, including those who do not use AI themselves. This involves understanding the technology, its social impact, and responsible use.
Advanced level, for those who actively use AI in their daily practice, for example when creating teaching materials. This includes additional skills, such as effective collaboration with an AI system and evaluating its outcomes.
Expert level, for those who build, adapt, or code AI systems themselves, or teach about them in the upper years of secondary education.
The AI Regulation requires all staff to have the basic level. Who needs to be at the advanced or expert level depends on the choices and ambitions of the school. However, Kennisnet emphasises the importance of having this knowledge within the team.
It matters, even without AI in the classroom
One of the strongest passages in the article addresses the idea, "I don’t use AI, so I don’t need to know anything about it." Kennisnet calls this understandable, but no longer tenable. After all, pupils do use AI, for homework, summaries, and assignments.
Teachers who do not use AI themselves must still be able to adapt their curriculum and assessment to a world with AI. And they must be able to help pupils reflect on their own use. This requires basic knowledge, even in a deliberately AI-free classroom.
What’s in the toolkit
In addition to the framework itself, Kennisnet has created a toolkit for schools working on AI literacy for the first time. With it, you can map out your team’s development needs and set priorities.
The complete framework with explanations and a compact version with all indicators per level on one page.
A self-evaluation for the basic level.
A presentation to guide a team session on AI literacy.
A document on the background and development, including the choices made in adapting AI-GO! to primary and secondary education.
The practical tips provided by Kennisnet are helpful. Measurement is not compulsory, certification is not required, and there are no mandatory training sessions. Internal registration of what you organise for whom is sufficient. On 11 November, Kennisnet will host a webinar on working with the framework and toolkit.
Where EduGPT connects
At CivAI, we are pleased with this framework, as it precisely describes what we see daily in practice with EduGPT. You do not learn AI literacy from a leaflet. You learn it by practising in a safe environment, making mistakes, and reflecting together. Especially in an era of AI, it is reassuring that there is now a shared language for what teachers need to know and be able to do.
We immediately recognise a number of indicators from the framework. At the basic level, Kennisnet asks that you can assess whether AI use aligns with values such as privacy, confidentiality, and data protection. That conversation is much easier if you work in an environment with European hosting, where the school retains ownership of the data and personal data can be anonymised.
The framework also mentions the environmental impact of AI as basic knowledge. In EduGPT, teachers see an indication of energy, water, and token usage per conversation. This way, an abstract concept becomes concrete in daily practice.
At the advanced level, it is about effectively directing AI, evaluating outcomes, and giving AI an appropriate place in teaching practice. This is at the heart of how we have designed EduGPT. Teachers compare two models side by side, see the differences in outcomes, and thus learn why you should always check AI output against your own subject knowledge.
With the Prompt Companion, teachers create their own AI Companions for their subject or their pupils. In our view, this is the best form of AI literacy. Teachers become designers of their own AI support rather than consumers of a tool, and the judgement of what constitutes good education remains with the teacher. Our founder Pascal Mariany calls this in his book The Art of Living Together with AI the Reasoned Responsibility: ownership of judgement remains with the human.
We are pleased that Kennisnet explicitly mentions coding at expert level. We will be happy to return to this later this autumn.
Getting started together
For schools wishing to start with the framework, we advise starting small. Do the self-evaluation with the team, organise a team session, and choose two or three indicators to really focus on in the coming half year. As Kennisnet itself says, don’t get overwhelmed and do what suits your school.
Would you like to practise with your team in a safe AI environment, have a team session facilitated, or discuss how the framework translates to your practice? Feel free to contact us at info@civai.eu. We are happy to think along with you.
Want to read more? View the article and toolkit on the Kennisnet website or download the compact framework (pdf) directly. You can find more about EduGPT at edugpt.nl.
Image for illustration, created with AI