This paper analyses how generative Artificial Intelligence (AI) systems can facilitate teaching and significantly improve students’ learning experiences. It compares them with technological support tools available before the advent of AI. Information and communication technology has provided tools to promote integration, including in the classroom, since the late 20th century. Today, Large Language Models (LLMs) – also called ‘conversational agents’ – represent a revolutionary opportunity to transform classrooms and enable the inclusive integration of linguistically and culturally diverse learners in educational settings. The present analysis was carried out in light of the European Union’s ethical principles for AI in education, which were used to test simple applications of ChatGPT in a classroom in a high school in Italy to explore the practical and theoretical frameworks in which AI supports the integration of students from different cultural and linguistic backgrounds. Based on the results of this simple case study, it is possible to highlight the pedagogical potential of LLMs in promoting inclusive, supportive and multicultural education, although they must be evaluated and identified. This paper also highlights the central role of AI in building equitable, student-centred classrooms. The paper also highlights some of the ethical issues that may arise from the use of AI systems in education.