Wikipedia and large language models: perfect pairing or perfect storm?

Abstract

Purpose The purpose of this paper is to explore the potential benefits and challenges of using large language models (LLMs) like ChatGPT to edit Wikipedia. Design/methodology/approach The first portion of this paper provides background about Wikipedia and LLMs, explicating briefly how each works. The paper's second section then explores both the ways that LLMs can be used to make Wikipedia a stronger site and the challenges that these technologies pose to Wikipedia editors. The paper's final section explores the implications for information professionals. Findings This paper argues that LLMs can be used to proofread Wikipedia articles, outline potential articles and generate usable Wikitext. The pitfalls include the technology's potential to generate text that is plagiarized or violates copyright, its tendency to produce “original research” and its tendency to generate incorrect or biased information. Originality/value While there has been limited discussion among Wikipedia editors about the use of LLMs when editing the site, hardly any scholarship has been given to how these models can impact Wikipedia's development and quality. This paper thus aims to fill this gap in knowledge by examining both the potential benefits and pitfalls of using LLMs on Wikipedia.