Currently reading:
**new preprint** on arXiv:
"ChatGPT “contamination”: estimating the prevalence of LLMs in the scholarly literature" by @generalising
@rmounce @generalising I wish this had mentioned covered second-language learners who may use ChatGPT to sound more natural in academic English.
I'm not sure how to test for that except to compare papers from individuals, though.
@richlitt @rmounce I would agree - I think I originally thought of that as covered by the ref to "copyediting", but V2 will try and tease out all those different (legitimate!) use cases a bit.
It's challenging though - since there's no disclosure of what/how the tools were used, we don't really have any way to tell what's causing these markers without really digging into individual cases.
