Content and Language Integrated Learning (CLIL) requires materials that are both linguistically accessible and conceptually accurate, yet producing such resources is time-consuming and commercially available options remain limited.Recent attention has turned to generative AI (GenAI) tools such as ChatGPT, but their tendency to hallucinate, lack of source attribution, and potential for plagiarism limit their classroom reliability.This article examines Google's NotebookLM, a GenAI tool that grounds outputs in user-uploaded sources, offering verifiable traceability not available on most other platforms.Using open-access texts from E-International Relations, NotebookLM was employed to generate B1-level CLIL readings, comprehension questions, and discussion tasks for a Politics and International Relations CLIL seminar.Findings indicate that NotebookLM effectively reduces teacher workload while retaining disciplinary vocabulary and providing transparent source attribution.Generated materials were largely accurate, with comprehension questions that were mostly answerable and pedagogically usable.At the same time, outputs reflected the orientation of uploaded sources, underscoring that no text is neutral and highlighting opportunities to foster students' critical awareness of framing.Challenges include the need to update sources for currency, review simplification for conceptual precision, and ensure balance in perspectives.Overall, NotebookLM demonstrates strong potential to support efficient, reliable, and critically informed CLIL material development while keeping educators in control of content.
Craig LUCAS (2026) studied this question.