Title: Language across disciplines in the age of AI and EMI

Name of presenter: Maria Kuteeva

Affiliation and academic title: Professor of English linguistics at Stockholm University

​​

Abstract:

This talk revisits two essential concepts, language and discipline. Looking across three dimensions of language – structural, socio-ideological, and interactional (Kuteeva, 2023) – I argue that large language models (LLMs), work quite well across different languages as codes, especially for majority languages. Their use is less straightforward with regard to specific disciplinary discourses. When it comes to authorial voice, stance, and stylistic nuance, the LLM output often requires human intervention. How is this variation connected to academic disciplines? Disciplines are not set in stone: objects of study, research methodologies, and contextual factors impact knowledge-making and literacy practices. Recent debates across various research communities call for an increase in knowledge perspectives. In this context, GenAI-supported tools hold potential to address systemic biases, e.g. between Global North and Global South (Nature Editorial, 2023). LLMs are indeed useful for facilitating access to so far occluded genres, for providing translations, and for acting as a sounding board in the process of iterative prompting and interactive refinement (e.g., Mollick, 2023). At the same time, as LLMs lack agency and knowledge of the local context, they may drive both language use and knowledge construction towards further homogeneity and reinforcement of existing biases (Kuteeva & Andersson, 2024). To address these limitations, I suggest ways to foster collaboration between EMI and EAP professionals to support disciplinary literacy practices and critical stance by drawing on their knowledge of socio-material contexts and multilingual repertoires.