Generative AI and Unequal Englishes in ELT: From Algorithmic Gatekeeping to Critical Algorithmic Language Awareness

Generative artificial intelligence (GenAI) increasingly shapes everyday judgements about English in language education by producing, recommending, evaluating, and normalizing forms presented as correct, natural, professional, intelligible, or credible. Viewed through Unequal Englishes (UE), this growing authority matters because GenAI operates within institutions where Englishes already carry unequal legitimacy and value. This conceptual article proposes Critical Algorithmic Language Awareness (CALA) as a pedagogical response to algorithmic linguistic gatekeeping. CALA develops learners’, teachers’, and institutions’ capacity to examine whose linguistic values models privilege and to make contextually defensible decisions about accepting, modifying, or resisting algorithmic recommendations. The article identifies four forms of gatekeeping that motivate CALA: algorithmic standardization, linguistic self-repair, pragmatic and cultural normalization, and evaluative surveillance. Together, these mechanisms show how GenAI may reproduce linguistic inequality while creating opportunities for critical inquiry. ELT therefore needs to govern algorithmic linguistic authority while preserving informed, situated human agency in teaching and assessment.