Unlocking ChatGPT's Potential to Promote Gender‐Inclusive Language: Comparative Analysis of Epicene Pronoun Use by Chinese EFL Learners and ChatGPT

ABSTRACT A global shift toward gender‐inclusive language has led to the wide use of singular they as the preferred epicene pronoun. However, many learners continue to rely predominantly on generic he and he or she . Given that ChatGPT is trained largely on datasets from English‐as‐a‐first‐language (L1) speakers, among whom singular they has been broadly embraced, it may serve as a useful resource for modeling gender‐inclusive pronoun use for learners. This study compared epicene pronoun use in essays written by Chinese English‐as‐a‐foreign‐language (EFL) learners versus ChatGPT, drawing on two corpora of essays ( n = 9864 each) produced on identical writing topics. Results showed that learners largely preferred generic he , despite a gradual increase in singular they compared with essays produced approximately a decade ago. In contrast, ChatGPT overwhelmingly employed singular they , but occasional usages of generic he were also observed. Additionally, learner usage was primarily conditioned by antecedent individuation (i.e., notional number), followed by gender‐stereotypical associations, whereas ChatGPT's residual variation appeared to be solely driven by gender‐stereotypical associations. Overall, the findings point to learners’ continued reliance on gendered pronouns and highlight ChatGPT's pedagogical potential to support the development of gender‐inclusive pronouns in EFL contexts, provided that learners are guided to engage critically with its outputs.