I have a preprint out estimating how many scholarly papers are written using chatGPT etc? I estimate upwards of 60k articles (>1% of global output) published in 2023. https://arxiv.org/abs/2403.16887
How can we identify this? Simple: there are certain words that LLMs love, and they suddenly start showing up *a lot* last year. Twice as many papers call something "intricate", big rises for "commendable" and "meticulous".
@sbszine yeah, I think it's very variable between document types - I found that some of the terms that were significant for peer reviews had little to no difference for papers. I guess it's down to the nuance of what terms are expected (by a human) in what contexts!
@generalising Ah, right! My use case is reading tender documents and a text search for 'moreover' is (anecdotally) reliable at finding LLM text.