Comment by m0rde
5 hours ago
I'm a bit confused. Someone posted fake stats online and saw ChatGPT visit their page. Then they immediately assume that ChatGPT is falling for those fake stats and using it in chat responses or analysis input? Where's the proof? How do you know it didn't search, find, and discard?
The author also sees their stats in a paper and a few other sites and assumes ChatGPT put it there? Not a human visiting the fake stats site?
Did I miss the evidence that shows ChatGPT is actually using the fake stats? Serious question.
Based on the article, it's because Google (and presumably other search engines) did not surface the website, only LLMs crawled it.
Even if some LLMs discarded it, only they were the ones returning the results to users. Occam's razor tells me that LLMs cited the website and humans were too lazy to check. That the stats ended up being cited is a lack of scrutiny on both the LLMs and the human that should have reviewed it.
If the end goal of LLMs are autonomous agents, i.e. removing human oversight, then this will be what proliferates on the web.
> Where's the proof? How do you know it didn't search, find, and discard?
They found their page being cited in various AI generated slop around the internet, including an acedemic paper, investment research, an Indian publication, etc.
These days I find it's my own fault to spend time reading anything on a .ai top domain.