The application of AI chatbots in language teaching and learning has created a new domain of opportunities for the ELT community. This study attempted to investigate the effectiveness of using AI chatbots in genre instruction in an EAP course aimed at improving TEFL MA students’ academic writing skills at Ilam University, Iran. To this end, following a proficiency test and a test in academic writing, 36 students enrolling in the course were allocated to two homogeneous classes of 18 with equal number of learners in the proficiency levels upper intermediate and advanced. Length, frequency and number of sessions in both classes were equal. In both classes, a collection of the abstracts of 60 research articles in applied linguistics and TEFL formed the sample corpus used for instruction. In one class, five abstracts were presented each session and the AI chatbots ChatGPT and Microsoft Copilot were queried and prompted to analyze the abstracts, highlight their points, rewrite them and review and evaluate the way information, data, arguments, methodology, findings and results were presented within the abstracts, and AI-generated contents and output produced by the chatbots were presented to the students. In the other class, five abstracts were presented each session with the instructor as the only source of analysis and insight on the abstracts. At the completion of the course, each student was assigned to write six research paper abstracts and a group of 12 raters rated the abstracts with a score of zero to 10 in a double-blind manner. The results of the ratings showed that students of the AI-assisted class outperformed the students of the other class. This can confirm the effectiveness of using AI-powered chatbots in genre instruction. Also, students with advanced level of proficiency achieved higher scores and demonstrated higher improvements compared to their entrance level. Further studies focusing on effectiveness and usefulness of each AI chatbot and the effects of, or the relationship between, different factors and parameters such as individual differences, age, academic subject, topic, focus of the courses and application of different AI-powered solutions in different stages and processes of instruction, teaching and learning in different educational settings are recommended.
