A duoethnography of recognition in AI-mediated academic writing

Generative artificial intelligence (AI) can produce fluent prose without revealing the writer’s underlying reasoning. We examine how two multilingual academics interpret the relationship between AI-mediated fluency and judgements of academic competence. A duoethnographic reanalysis of a six-month inquiry draws on reflective accounts and dialogue alongside selected ChatGPT traces. Four curated scenes explore reported praise, relief from writing difficulty, the normalisation of a professional voice, and contrasting responses to fluency. Reciprocal questioning and comparison with material that challenged our initial explanations informed the analysis. Our interpretations concern experienced and anticipated judgements. We develop relational literacy as the situated capacity to interpret and take responsibility for the relations through which AI-mediated writing is produced and judged. Connecting AI literacy with academic literacies, we examine how textual qualities become grounds for attributing competence. We propose teaching and assessment activities that make composing decisions discussable without treating polished prose as sufficient evidence of learning.