Teaching AI Literacy in diverse Classrooms

By Ans De Nolf

2026/05/15

Over the past few months, our AI literacy lesson travelled across 18 classrooms in a Belgian bridging programme for newly arrived non-Dutch-speaking students (OKAN). These students are still learning the language while simultaneously finding their way in a new school system, and often a new life. This makes the classroom a uniquely dynamic space, where language learning, adaptation, and digital practices intersect, and where questions of AI literacy become both urgent and complex. The pupils of OKAN are a highly diverse group: from young beginners (starting around age 12) with limited knowledge of Dutch to older students with relatively advanced language and digital skills.

We started our class by asking who knows or uses AI, and a pattern emerged almost immediately: many students were already using AI. Tools such as ChatGPT, Gemini, Microsoft Copilot, and Siri came up in nearly every class, often as part of students’ daily routines for translation or homework support. Yet, we also know that frequent interaction with AI does not necessarily imply understanding how it works or how to engage with it critically.

What stood out?

Across classrooms, familiarity with AI was widespread but uneven. In almost every group, most students had at least some experience using AI, yet the range was striking: some demonstrated quite advanced knowledge, while others were encountering these tools for the first time.

A small number of students also took a more critical stance. One student, for instance, explicitly not using AI because she saw it as conflicting with her identity as an artist. More commonly, however, AI was approached pragmatically, primarily as a tool for translation and school support.

Engagement patterns were equally telling. Students were consistently more involved during hands-on activities than during more explanatory moments, suggesting that active experimentation plays a key role in how they make sense of AI.

Picture: Students working in groups on the assignment

What did we learn?

1. Language is the biggest challenge

Many students struggled with the questionnaire and concepts, even when translated. We learned to use simpler language, visuals and examples, instead of long explanations.

2. Students learn best by doing

The creative exercise, modifying paintings with AI, proved particularly effective. It generated immediate enthusiasm: students engaged actively, experimented creatively, and took clear pride in their outcomes. At the same time, the activity opened up space for more substantive discussions, including questions around privacy, copyright, and the status of AI-generated or manipulated images.

3. Group work helps a lot

Students supported each other with language and ideas. But this only worked well when enough teachers or helpers were present.

4. Teachers want support

A cross schools, teachers consistently noted the absence of structured AI lessons, despite the fact that students are already engaging with these tools. This gap was widely recognized, and many teachers expressed interest in accessing and using our materials in their own classrooms.

5. Practical issues matter

Logging in, blocked websites, and technical problems often slowed us down. Good preparation is very important.

Teaching AI in OKAN classes works best when it is simple, visual and hands-on. Students are curious and enjoy trying things out, but they need clear explanations and practical activities to really understand.

It is also not just about showing them how to use AI. It is just as important to talk about when to use it, what not to share, and how to think about the answers it gives. The goal is to help students feel confident using AI smartly and safely, not just a quick way.

Key takeaways

Access is no longer the main issue. Understanding is. Without targeted support, existing inequalities risk being reproduced in how students use and are shaped by AI.

If students learn AI best by doing, then teaching AI literacy means more than explaining tools: it means creating spaces to experiment, question and reflect. That is where real understanding begins.