AI in the Box: What (Not) Is Intelligence?

As part of the AITECH project activities, Dino Nejašmić presented the talk “AI in the Box: What (Not) Is Intelligence?” at the CUC 2026 conference in Rovinj (April 15–17, 2026).

Artificial intelligence is increasingly used in education, yet it is often treated as a “black box” — a system that produces answers without sufficient understanding of how those answers are generated. This approach can lead to superficial learning, uncritical acceptance of results, and overestimation of AI capabilities.

This presentation introduces the concept of “AI in the box” as a way to better understand the role of artificial intelligence in education. It explains the basic principles behind large language models (LLMs), emphasizing that their outputs are primarily based on statistical patterns and probabilities rather than true understanding.

Special attention is given to different levels of transparency in AI systems — black box, grey box, and white box — and their implications for teaching and learning.

The final part of the session presents a practical workshop concept designed for students, aimed at explaining AI in a simple and accessible way while encouraging critical thinking about its use.

This activity was supported by the project “Artificial Intelligence for Future Technical Education and Industrial Competitiveness (AITECH), IP-UNIST-48”, funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.