Reconstructing stance in EFL doctoral thesis writing through generative artificial intelligence

Since its emergence in 2022, Generative Artificial Intelligence (Generative AI) has rapidly gained widespread application in the field of academic English writing (EAW), triggering multifaceted transformations in EAW practices, particularly in the construction of authorial stance. Doctoral dissertations, due to their stringent demands for academic depth and originality, serve as an ideal subject for investigating the impact of Generative AI. EFL (English as a Foreign Language) doctoral students, often prone to language anxiety, are more likely to adopt passive writing strategies, which may undermine the expression of authorial stance. Therefore, this study, based on corpora from sub-disciplines within applied linguistics with high and low exposure to AI, employs a Difference-in-Differences (DiD) design to systematically examine the influence of Generative AI on stance construction in doctoral dissertations during two phases: the introduction period (2019–2021) and the diffusion period (2022–2024). The findings indicate that Generative AI is reshaping stance strategies in academic writing, driving a shift towards more deterministic and objective argumentation styles, while potentially weakening the explicit expression of authorial identity.