AI as a tool for the student in PBL and modeling – empirical validation of a didactic model for active ai learning in technological education

Abstract

The integration of artificial intelligence (AI) into secondary education is increasingly emerging as a strategic priority, but existing approaches often focus on theoretical knowledge about AI rather than its active application. The model – "AI as a Tool for the Student (PBL and Modeling)" – has been described in a collective monograph as a theoretical didactic framework. The present article presents it not as a repetition but as an enhancement with empirical value. Here, the model is not only described but also tested, implemented, and evaluated in a real school environment. It transforms AI from an automation tool into an instrument for competency acquisition in symbiosis with the STEM approach. The model is operationalized through four sequential phases: problem definition, data collection, AI analysis, and modeling/solution. Methodological guidelines for implementation are derived from a pilot study conducted at the Varna Maritime High School "Saint Nicholas the Wonderworker" (VMG) , within the subject "Fundamentals of Artificial Intelligence" for students majoring in Computer Hardware and Technologies. The pilot implementation integrated IoT sensors for temperature and humidity (ESP8266 with DHT22, Unihiker K10) with a MIND Link connection to an AI model. Results from 48 participating students (12 teams) showed an average ML model accuracy of 87%, 94% student engagement, and a 62% self-reported improvement in applied AI skills (paired t-test, p < 0.001). The article concludes that the model is not only theoretically described but has now been empirically validated. It successfully develops applied AI competencies, stimulates critical thinking, and prepares students for professional careers in AI, IoT, and data science. These results have been achieved provided that systematic teacher training and institutional support are ensured.