Taking a Stance: Scholar‐Written and ChatGPT‐Generated Academic Blogs

ABSTRACT The advent of Large Language Models in the last few years has dramatically changed the landscape of academic writing, not least in the domain of scientific blogging. Not only can GenAI tools draft posts quickly, but they can potentially synthesise and transform complex research articles for a blog audience of lay readers. In this way they can bridge the gap between academia and the public while at the same time assisting non‐native speakers in producing fluent academic writing. Despite this development, however, little research has systematically explored the rhetorical construction of these AI‐generated blogs and how they compare with those created by human scholars. In this paper, we explore this comparison by focusing on stance. Using Hyland's (2005) stance model, we investigate similarities and differences between scholar‐written blogs on COVID‐19 and those generated by ChatGPT‐5 from the same 100 research articles. Our findings highlight how stance functions as a key resource in the reworking of scientific knowledge for public audiences and contribute to ongoing debates about authorial voice, knowledge dissemination and accountability in AI‐generated academic writing.