Writing-With ChatGPT Across Uneven Terrains: A Gendered Duoethnography in Vietnam and Belgium

This article examines how GenAI reshapes academic writing when encountered from uneven institutional, linguistic, and gendered positions. Drawing on a six-month duoethnography, we trace our Writing-With ChatGPT as two Vietnamese doctoral students in contrasting settings: a public university in Belgium and a doctoral program at a public university in Vietnam, while employed as a lecturer at a private university in the same region. Our data corpus includes co-written dialogues, ChatGPT transcripts, reflective journals, and voice memos in Vietnamese and English, which we analyze through the lens of Writing-With as a relational, recursive stance. We map our experiences across four entangled dimensions of what we call institutional positioning (not a geographic location): (1) infrastructural conditions (bandwidth, devices, platform access), (2) institutional policy (permissiveness, assessment regimes, enforcement), (3) discursive legitimacy (expectations for “standard” academic English, journal norms), and (4) geopolitical positioning (citizenship, mobility, visa regimes). Against this backdrop, we read our scenes through a spiral of seven moments, including pausing, disrupting, surrendering, doubting, reintroducing, accepting, and continuing, that emerged from our Writing-With process. Our analysis shows that ChatGPT’s algorithmic “smoothing” of prose is experienced as amplification for one author and erasure for the other, that gendered role socialization shapes comfort with GenAI and what counts as responsible experimentation, and that institutional ambiguity around GenAI usage redistributes ethical risk unevenly. We argue for understanding ChatGPT as a partial, situated interlocutor whose affordances and harms are co-produced with infrastructural arrangements and disciplinary standards.