Machine-Translating Feminism: Generative AI, Epistemic Injustice, and the Politics of Gendered Meaning in Indonesian Feminist Discourse

In transnational feminist knowledge circuits, generative AI has become a decisive yet insufficiently theorised mediator of gendered meaning across languages. Large language models do not simply transfer words between Indonesian and English; they set the moral, political, and gendered conditions under which feminist vocabularies become intelligible. This article intervenes at the intersection of feminist translation studies, epistemic injustice theory, and multilingual AI critique to theorise what we term machine-translating feminism: the process through which generative AI translates feminist language while reformatting feminism into a more manageable discursive object. Departing from English-centric AI bias scholarship, we turn to Indonesian feminist discourse as a Global South site where womanhood, sexuality, violence, and anti-feminist backlash are mediated through contested vocabularies. Through close readings of Ayu Utami’s Saman and Leila S. Chudori’s Laut Bercerita, alongside public feminist vocabularies, the article identifies four diagnostic patterns of translational distortion: lexical neutralisation, moral simplification, depoliticisation of structural violence, and patriarchal overcoding. The analysis shows how generative AI operates as a translation regime rather than a neutral tool, producing what we term patriarchal legibility: a mode of readability in which feminist conflict becomes internationally consumable by becoming less threatening to patriarchal, institutional, and Anglophone interpretive regimes.