This should be cause for contempt. This isn’t much worse, IMO, than a legal briefing mentioning, “as affirmed in the case of Pee-pee v.s Poo-poo.” They’re basically taking a shit on the process by not verifying their arguments.
Can we normalize not calling them hallucinations? They’re not hallucinations. They are fabrications; lies. We should not be romanticizing a robot lying to us.
Pretty engrained vocabulary at this point. Lies implies intent. I would have preferred “errors”
Also, for the record, this is the most dystopian headline I’ve come across to date.
If a human does not know an answer to a question, yet they make some shit up instead of saying “I don’t know”, what would you call that?
that’s a lie. They knowingly made something up. The AI doesn’t know what it’s saying so it’s not lying. “Hallucinating” isn’t a perfect word but it’s much more accurate than “lying.”
“Fabrications” seems ok to me
Bullshit.
This is what I’ve been calling it. Not as a pejorative term, just descriptive. It has no concept of truth or not-truth, it just tells good-sounding stories. It’s just bullshitting. It’s a bullshit engine.
How is this not considered fraud? Or at least hold them in contempt.
There are so many services for formatting sources…why use an LLM for this?
Vibe litigation
This doesn’t appear as bad as some of the other ai legal stuff. Formatting references isn’t really about generating content as much as structuring it and AI (usually) doesn’t have the kind of problems with hallucinations when just tasked with reorganizing data. I’ve used GPT for reformatting references to APA style and it worked really well. I’m surprised Claude couldn’t handle this task.
Also bummed that there doesn’t appear to be a book called a statisticians guide to making inferences with noisy data, because that sounds like a book worth checking out.
They gave it a link to the paper, not the text of the paper. So it probably couldn’t actually access the URL and just pulled from its training.
