BRAIN IN A JAR: INTERACTIVE EDUCATIONAL TOOLS FOR CRITICAL AI LITERACY
University of Split, Faculty of Science (CROATIA)
About this paper:
Conference name: 18th International Conference on Education and New Learning Technologies
Dates: 29 June-1 July, 2026
Location: Palma, Spain
Abstract:
Children increasingly meet artificial intelligence (AI) in everyday apps and devices, yet their explanations often rely on “black-box” thinking or human-like assumptions. Brain in a Jar is an interactive workshop for primary school learners (ages 10–12) that helps students examine AI claims with a cautious, evidence-based mindset. The workshop centers on a physical “smart box” with an input and output slot. One student takes the role of the box operator and, out of the group’s view, uses rule sheets placed inside the box to map each input to a corresponding output. This setup lets the class observe consistent behaviour first, and then unpack how it was produced.
Two task variants support different discussions. In a number-lookup variant, students submit prompts that appear difficult to solve quickly and the operator returns the pre-prepared matching answer. In a symbol-mapping variant inspired by the Chinese Room, students submit simple symbol pairs and the operator applies explicit rewriting rules to produce the output. The activity is followed by a structured discussion where students must justify their claims. They describe the steps the system followed, separate correct outputs from understanding, and propose what evidence would be needed to say a system understands. A final reflection prompt asks students to state one thing the system can do and one thing it cannot, encouraging them to distinguish between capabilities and limitations.
The workshop also uses a brief robot demonstration to show the same mapping idea in a different setting. Students show printed markers to a humanoid robot, which identifies the marker and replies with a predefined response. The facilitator then links the robot’s behaviour back to the same kind of input–rule–output process used in the smart box, and steers discussion away from novelty toward what the system is actually doing.
Building on our previously reported classroom pilots, we focus here on facilitation moves that keep attention on inputs, rules, and the limits of rule-following, especially when using an optional robot demonstration. It provides ready-to-use materials, timing, and discussion prompts for classroom teaching or outreach, plus simple alternatives when a robot is not available (video or role-play).Keywords:
AI literacy, critical thinking, interactive learning, classroom workshop, human-AI interaction (HAI), workshop design.