This study investigates how English for Academic Purposes (EAP) practitioners in British higher education institutions use Generative AI (GenAI) tools for classroom materials development. Drawing on semi-structured interviews and AI chat logs from 13 qualified and experienced EAP practitioners, the article explores their motivations, engagement strategies, evaluation practices and ethical considerations when using tools such as ChatGPT and Microsoft Copilot. Framed by Moorhouse et al.'s (2024) Professional GenAI Competence model, findings reveal that EAP practitioners are motivated primarily by efficiency gains and personalisation opportunities, while also valuing GenAI as a collaborative partner for idea generation and materials refinement. Participants employed differing prompting strategies - from detailed specification to minimal dialogic interaction - demonstrating varying approaches to technology's affordances. Most participants engaged in iterative revision processes guided by pedagogical knowledge to align outputs with learning objectives. While concerns about content accuracy and bias were acknowledged, these did not significantly impede the adoption of GenAI tools. More prominent were ethical considerations around transparency and authorship attribution, influenced by unclear institutional guidelines and fears of professional judgment from colleagues and students. A notable finding is GenAI's potential to democratise materials development, enabling practitioners to create professional-quality resources independently of commercial publishers. These insights inform recommendations for targeted teacher training and institutional guideline development to support effective and ethical GenAI integration in EAP contexts. • EAP practitioners use GenAI primarily for creating texts for genre analysis. • Efficiency, personalisation and lack of resources motivate GenAI adoption. • Many EAP practitioners value GenAI as a dialogic partner for materials refinement. • Ethical concerns focus on transparency and authorship rather than content accuracy. • GenAI democratises materials design, reducing reliance on commercial publishers.
Ngo et al. (Thu,) studied this question.