NO AI ACT.

Teachers · Lab · 30/60/90 minutes

The AI Act classroom lab

Everything you need to turn the game and this site's materials into a lab you can run tomorrow: three paths of increasing length, the printable materials, the pitfalls to watch, and an honest debrief on what to expect.

01What you need

0230-minute path — one case and a debrief

  • 5' — Hook: "What would happen if no AI rules existed at all?" Collect three hypotheses on the board.
  • 15' — Play: each pair opens one case (start with the scoring city) and reaches the report.
  • 10' — Debrief: compare two different classifications that emerged; close with the banned / high-risk / transparency distinction.

0360-minute path — case + worksheet + rubric

  • 10' — Hook and a mini-introduction to the risk-based approach (risk categories).
  • 20' — Play: one case in pairs, with the task of noting the decisive piece of evidence.
  • 15' — Worksheet: each pair fills in the risk-classification worksheet on the played case (or on a real system you choose).
  • 15' — Rubric-guided discussion: use of evidence, regulatory language, affected people, contestability.

0490-minute path — a full lesson

Use the introductory lesson plan as the backbone (scenario, risk-based approach, game, reflection) and slot the worksheet and rubric into the final phase. Alternatively, build two of the five ready-made activities around one central case. Over several sessions: one thematic chapter per lesson (prohibited practices → high risk → transparency) — see also the other lesson plans.

05What to expect (honestly)

The lab trains classification, use of evidence and reasoned discussion — but the game's educational effectiveness has not yet been empirically validated: treat it as promising teaching material to evaluate with your own instruments, not as an intervention with demonstrated effects. Classroom assessment stays formative (the rubric is not a grade). The full picture is in research and methodology.

Try the game Printable materials

06Official sources and transparency