Le IA generative per il decision making. Attendibilità degli LLMs a supporto di decisioni di natura didattica ed educativa

This article presents an experiment with 140 teachers and educators in ini-tial training at the University of Bologna aimed at assessing the reliability of Artificial Intelligence systems to support the selections of applications, also AI, for teaching (Panciroli – Rivoltella, 2023). The research involves three levels of analysis: 1) the mapping of available applications through the use of three specialised search engines (Futurpedia, Ailocater and GpteAi); 2) the interaction with Chatbots, to select within this mapping, the specific ap-plications deemed most functional by chatbots himself (Manovich, 2020); 3) the validation of the choices suggested by Chatbots through a testing activity carried out by teachers and educators to assess their technical reliability to support decision making in educational systems. The expected outcome is a human-machine cooperation model for decision making and instructional design.