XClose

Advanced Research Computing

Home
Menu

TechSocial Series - January 2025

ChatGPT “contamination”: estimating the prevalence of LLMs in the scholarly literature', by Andrew Gray, UCL's very own Librarian

Speaker Bio

Andrew Gray is a librarian at UCL with a specialisation in bibliometrics, the analysis of large-scale publication patterns and citation networks. Over the last year he has been investigating the ways in which AI tools interact with scholarly publishing and searching.

Other works: Chatbots Have Thoroughly Infiltrated Scientific Publishing (Scientific American)


Abstract

AI-derived text is increasingly common in scholarly publications, with a significant amount of papers in some fields now showing evidence of having used large language models in some way, and the research community divided on how best to handle the situation. This talk will look at some recent research which tries to identify how widespread the use of large language models is, and discuss what the rise in AI-generated text might mean for the integrity and quality of scholarly communications in the future.

Fancy some 'light', pre-talk reading? Then check out Andrews publication!

 


Archived Media

Recording | Slide Deck