RECOMMENDED: Large Language Models and Academic Writing

The South African Journal of Science recently published an article by Martin Bekker (University of the Witwatersrand) that explores a tiered model for assessing academic authors’ engagement with large language models (LLMs) like OpenAI’s ChatGPT.

“Large language models and academic writing: Five tiers of engagement” offers guidance for academic journal editors, university instructors and curriculum developers (and library workers) on thinking about the different modes of authorial engagement with LLMs for academic writing. The article proposes a five-tier system “to simplify thinking around permissions and prohibitions related to using LLMs for academic writing. While representing increasing ‘levels’ of LLM support that progress along a seeming continuum, the tiers in fact represent paradigmatically different types of mental undertakings” (p. 2).

The tiers include 1: Use ban, 2: Proofing tool, 3: Copyediting tool, 4: Drafting consultant, and 5: No limits. Bekker proposes an ethical framework for evaluating potential harms and benefits for authors’ use of LLMs at each tier of engagement. Concluding with a brief discussion of “AI hype and despair,” this paper makes an interesting contribution to the ongoing conversations in higher education across the globe around emerging AI technology’s use and impact on academic publishing.

Read the full open-access article on the publisher’s website.