Large Language Model AI and Lx Academic Writing: Challenges and Opportunities
Large Language Model AI and Lx Academic Writing: Challenges and Opportunities

Author(s): Oliver James Ballance, Averil Coxhead
Subject(s): Foreign languages learning, Computational linguistics, Stylistics
Published by: Wydawnictwa Uniwersytetu Warszawskiego
Keywords: computer-assisted writing; computer-assisted language learning; academic writing; artificial intelligence; large language models
Summary/Abstract: This paper discusses the implications of large language model artificial intelligence for teaching, learning, testing, and producing Lx academic writing. The paper first provides a brief, non-technical description of artificial intelligence, machine learning, deep learning neural networks, and large language models (LLMs), drawing out the applications of the latter most relevant to the field of Lx academic writing. It then discusses and exemplifies the relevance of LLMs to influential models of academic writing, as outlined by Hyland (2019): structural, functional, creative expression, process, content, and genre-focused writing pedagogies. The next section discusses six major issues raised by the emergence of LLMs: differences in language support, the prevalence of LLM hallucinations, lack of intentionality, redefinition of the knowledge Lx writers need to bring to the writing task, assessment issues, and ethical issues. The paper concludes with recommendations as to how this technology can be incorporated within a consistent and defensible pedagogy, the research agenda needed to support such a pedagogy, and the need for a revaluation of the underpinnings of academic writing that the emergence of this technology necessitates.

• Details
• Contents