Digital linguistic imperialism: can we decolonize English when the AI writing tools are colonial?

The integration of artificial intelligence (AI) tools into language learning presents a paradox: while promising democratization, they often reinforce linguistic hegemony. Through a methodological innovation, this study employed a 12-week critical post-human ethnography in an Iranian English as a Foreign Language (EFL) context, combining algorithmic audits of Grammarly, QuillBot, and ChatGPT with classroom observations of 23 MA EFL advanced university students (including screen recordings, interviews, and think-aloud protocols) and a counter-infrastructure co-design workshop. Drawing primarily on frameworks of infrastructural violence and agential realism, with data colonialism and a contextualized understanding of raciolinguistic ideology serving as contextual lenses, results indicate AI tools function as colonial infrastructures that systematically pathologize non-Inner Circle linguistic features, eliciting responses ranging from linguistic shame to defiance. Crucially, co-designing a localized micro-model validated translingual identities and enabled infrastructural sovereignty. The study contributes a novel methodological framework for auditing digital linguistic imperialism and proposes a decolonial post-human pedagogy centered on infrastructural co-governance.