The present study explores opportunities and risks offered by generative AI chatbots in the educational context. The focus is on examining the effective ability of professional and expert educators to distinguish between human-written and AI-generated argumentative essays, utilizing a methodology inspired by Alan Turing’s imitation game and, less directly, by a recent application to a group of German teachers. Data collected with an anonymous questionnaire from 180 Italian teachers reveal difficulties in identifying AI-generated texts, with the accuracy influenced by factors such as familiarity with AI tools and technological expertise. The findings highlight that AI-generated texts have some common linguistic characteristics, such as clarity, lexical variety, and structural logic, though human-written texts achieve better results in expressiveness. The study underscores the necessity for critical AI literacy among educators and their students to propose innovative strategies for integrating AI into education while addressing the main ethical concerns.
