Young learners and AI: Deciding when to bot.

The AI-powered technologies, and chatbots in particular, are increasingly promoted as solutions to enduring challenges in English language teaching (ELT). While these tools are often praised for their capacity to deliver interactive and personalised learning experiences alongside traditional instruction, their rapid uptake in educational contexts has outpaced critical reflection on pedagogical readiness. The growing imperative to adopt AI raises fundamental questions about whether teachers and learners are sufficiently prepared - linguistically, cognitively, and emotionally - to engage productively with such technologies. Rather than assuming that earlier exposure is inherently beneficial, this study interrogates when and how AI should be introduced to support meaningful language learning. Drawing on an on-site empirical investigation, the study examines the use of AI in English-medium Content and Language Integrated Learning (CLIL) classrooms, focusing on history and art in a bilingual school in Rome, Italy. It analyses interactions between primary (Grade 5) and lower secondary (Grades 6–7) students and AI-based tutoring tools on the SchoolAI platform. Through qualitative analysis of student–chatbot exchanges, the research critically explores how learners negotiate meaning with AI, revealing both the pedagogical affordances and the constraints of AI-mediated interaction in ELT settings. The presentation situates the study within current debates on AI in language education, challenging techno-optimistic narratives by foregrounding classroom realities. It outlines the research methodology and key findings, supported by illustrative excerpts from student-AI interactions, and concludes with a critical reflection on the implications for curriculum design and teacher decision-making