Research methods and generative artificial intelligence in applied linguistics

Since the release of ChatGPT, there has been an explosion of interest in Generative Artificial Intelligence (GenAI) and a desire to understand how these tools can be utilised in almost every domain of human activity.Never before in human history have we had a tool that could simulate so many human capabilities.Yet, just as with humans, these tools have been found to have substantial limitations.These limitations have led us to question the roles they can play in high-stakes tasks, such as research.At the same time, they have challenged our current conventions and norms of authorship, transparency, and accountability, thus forcing us to consider:• What tasks should be left to humans exclusively?• What tasks can AI do alone?• What tasks can/should AI and humans do together?For some scholars, clear red lines have been drawn.For example, 416 experienced qualitative researchers from 38 countries wrote a commentary rejecting the use of GenAI for reflexive qualitative research (Jowsey, Braun, Clark et al., 2025).They provided three legitimate reasons for their rejection: (1) GenAI as simulated intelligence is incapable of meaning-making; (2) Qualitative research should remain a distinctly human practice; (3) the established harms of GenAI, especially to the environment and workers in the Global South.Nguyen and Welch (2025) have similar concerns and arguments against the use of GenAI in qualitative research.For other scholars, GenAI offers significant potential to reshape "the academic research lifecycle-from ideation and literature discovery to hypothesis formation, methodological planning, and data acquisition" (Haber et al., 2025, p. 27).Meanwhile, other scholars, including us, take a pragmatic stance, trying to balance the potential of these tools with the explicit threats they pose to our epistemologies, methodological rigor, integrity, and ethics (Roe, 2025;Moorhouse, Nejadghanbar & Yeo, 2025;Moorhouse, Consoli & Curle, 2025).The current landscape is confusing and complex for researchers, readers, funders, practitioners, and the general public.We have seen many high-profile cases of uncritical and unethical use of GenAI in various aspects of research, such as, automated literature reviews without human oversight, the publication of AI-generated images without acknowledgement, and arguments supported by AIgenerated fictitious references (Gulumbe, 2024), leading to the contamination of our knowledge base with 'AI research slop', and erosion of trust in our research (Szudarski, 2025; Moorhouse & Nejadghanbar, 2025).Journals, research institutions, and journal editors have been found to have inconsistent guidelines, policies, and beliefs about GenAI in research and publication, creating opaqueness about acceptable and unacceptable uses of GenAI (See Yin & Chapelle, 2025, in this special issue).Some journals now offer GenAI-assisted pre-submission screening features that verify whether manuscripts