Chat GPT has a strong foothold on modern science papers. And it is getting worse by the hour.
Want to see:
Go to Google Scholar search, and type in. "certainly, here is"
You might be surprised.
/nosanitize
https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=%22certainly%2C+here+is%22&btnG=
These folks, or bots, or whatever, aren't even trying hard to edit, or disguise what they're doing.
So can you imagine how many there are that do light editing? Many, many more.
@corlin exactly
It isn't that hard or even time consuming to edit they phrases out I'd recon the "papers" that don't are deffo "bots" being used to auto generate them and maybe post them as "papers"
And like you say a lot will be getting edited as well taking they phrases out - there are tools though available that will determine if something has been AI generated - Tools like Copyleaks and AI Text Classifier are over 90% accurate and have proven fruitful in identifying AI-generated content
@corlin yeah it's deffo a big issue
Just posted a few articles about it and how a lot of Chat-GPT generated papers are showing up with "as of my last knowledge update" and "I don't have access to real-time data" in them from reputable and sketchy sources