There is a history of research on academic writing that connects the complex ways that writers think through writing processes to forge academic identities (Brandt, 2001; Ivanic, 1997). These studies of academic writing closely inspect practices like brainstorming, reflecting, and revising writing through observations in classroom environments and interviews to deepen an understanding about what it means to be an academic writer. Traditionally, academic writing has been examined as a steady process that requires time in institutional contexts with students honing the craft of writing for academic purposes (Thomson, 2023; Thomson & Kamler, 2016). However, in the age of AI, this picture has changed dramatically, with Large Language Models (LLMs) expediting academic socialisation for emerging writers and researchers through generated text that is prompted and customised to a writer’s needs and discipline. LLMs give developing academic writers immediate, interpreted text without the emotional and linguistic labour it takes to bridge cultures, linguistic repertoires, and registers. There are therefore more possibilities and affordances than ever to speed up an academic writing voice. By simulating languages and brokering across linguistic divides, LLM software gives rapid-fire cross-cultural understandings without the time and experience previously needed to navigate new languages and unfamiliar cultures. The core argument of this paper is that ChatGPT and LLMs more generally are social contexts where international students improvise academic identities and forge conceptions of writing. Key questions asked in the paper are: How do PGR international students perform academic identities and experience academic socialisation in the age of AI? In what ways is a writerly voice forged across AI platforms? This does not mean that AI-led writing is an academic panacea or in some way bypasses socialisation, rather AI must be seen as a co-traveller with international students as they navigate developing academic identities. 

The paper features a qualitative research study with ten international postgraduate (PGR) students at a university in the north of England. Modifying Gillen et al’s “A Day in the Life” (2007) method of data collection combined with focus group interviews, we asked ten PGR students to share their GenAI-led writing practices over one week which they then reflected on during focus groups. Interpreting their AI-led written artifacts as representations of thoughts and beliefs gave us a window into the social practices of academic reading, writing, and thinking that helped PGR students to forge academic identities. The paper begins with a presentation of Ivanic’s classic study of academic writing (1998), then presents Gillen et al’s “A Day in the Life” method (2007) to excavate digital writing practices, followed by key findings from our research study and the broader implications of it for ever-emerging AI research and scholarship.

A key finding in the paper is that the presence of AI appears to have triggered a reconceptualisation of authorship. Traditionally, Roz Ivanič (1998) framed the authorial selves through two primary forms of ownership, namely, the ownership of content and ownership of language (i.e. identification with the discoursal choices made). However, the data suggests a shift away from this binary toward a new focus on the writing process itself.

Gillen, J., Cameron, A., Tapanya, S., & Pinto, G. (2007). A Day in the Life: advancing a methodology for the cultural study of development and learning in early childhood. Early Childhood and Care. 177: 1-12. Ivanič, R. (1998). Writing and identity: The discoursal construction of identity in academic writing. John Benjamins Publishing Company. Thomson, P. & Kamler, B. (2016). Detox your writing. strategies for doctoral researchers. London: Routledge. Thomson, P. (2023). Refining your academic writing. Strategies for reading, revising, and rewriting. London: Routledge