"Certainly, here is a possible introduction for your topic" in scientific paper | Hacker News Reader

"Certainly, here is a possible introduction for your topic" in scientific paper

sciencedirect.com

74 pointsajdude13 comments

Certainly, here's a comment:

It seems like you inadvertently left in a placeholder or instruction for ChatGPT within your article's introduction. While it's an understandable oversight, it's essential to thoroughly review your text before publishing to ensure such remnants are removed. This helps maintain the professionalism and polish of your work. If you need further assistance or editing help, feel free to ask!

Scientific publishing is a total joke these days, especially for theoretical works (vs practical/experimental results).

We need a better way to sift through the mountains of make-work papers to find the stuff that actually works and is useful. Ironically, AI will probably help with that, at least until it outperforms humans at scientific discoveries itself.

I think the really disturbing aspect to me is that no editor or reviewer flagged this. It's literally the first sentence of the paper! Did they just read the abstract and say "meh, good enough"? Were they all using AI to review the paper?

Related Lemmy discussion of a similar incident from this week: https://lemmy.world/post/13128991

The relevant screenshot of the post: https://mander.xyz/pictrs/image/9d46f584-67a3-489d-92a5-c322... [An introduction of a published paper, containing the sentence "I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model."]

The feedback loop has begun, next generation models will be trained on previous models outputs.

Chat GPT prompt response? in the start of the second text block.

Sure seems likely! Definitely not natural flow to start a section lol

I'm unable to see the whole paper (although I can see the first sentence). Do the citations in that section actually cite relevant articles?

Amazing!

Their citations are going to go through the roof!

The journal editor should run a language model to detect if the paper includes some placeholder or instruction for a language model. Unfortunately they did not.

It's not even buried in the middle!

And the rest of it is predictably pretty crap. I'm no subject matter expert, but it doesn't sound like it's written by one (or any academic) either, even to me.

Whoops!

Reply on news.ycombinator.com