You're not tired of Twitter disputes yet, art you? Good!
There was a previous one (I saw it May 18th, but it was from a May 15th article) that had really gotten under my skin.
There are plenty of good reasons to be against AI and there has been a lot of that in the news lately:
The news sites seem to have been mainly driven by financial issues, and the Pope by concern for humanity. What set off the storm I noticed was caused by arXiv taking action against hallucinated citations:
https://byteiota.com/arxiv-bans-authors-1-year-for-ai-hallucinated-citations/
I had never heard of arXiv before, but they are "a free distribution service and open-access archive" focusing more on science and math.
Their repository includes economics, which was interesting because those were the hit dogs that howled loudest, so to speak.
Hallucinated citations have increased over the past few years. Rightly concerned about this, they have announced that authors submitting papers with references to non-existent papers will be banned for a year and that after their papers will need to be approved by a peer-reviewed source before acceptance.
As is frequently the case, the reason it became a thing is that someone upset said something that other people could not believe.
I believe that was started by James Miller, but then continued by Lenka Zdeborova (whom I'll get to later):
https://x.com/JimDMiller/status/2056514276311916768
You can read if you want.
Miller makes what almost seems like a reasonable point; if you spend a lot of time looking at citations, you know that errors happen.
That is true, but there is a big difference between, say, misspelling one author's name or getting a page number wrong, compared to referring to a paper that doesn't exist. An error could make referring to the original paper more difficult, but unless everything is wrong you can probably do some keyword searches referencing the authors or check a different year but the same month of the publication and find it, then check the reference.
His excuse is that he works with lots of coauthors, sometimes who speak different languages, and he has to trust them. When you start referring to twenty coauthors, I start to wonder what level of academic rigor is being upheld.
For the research I have been working on, the greatest number of authors on any of the papers in my literature study was six.
Some of this seems to relate to working with graduate students, and also trying to be cited as much as possible.
The graduate students probably need more supervision than putting twenty on a single paper would provide.
For being cited a lot, one of the things that came up was one person bragging about his H index, which started a big conversation on how some of the most important foundational work doesn't even get cited like that because it is part of the foundation and doesn't need a citation; everyone is familiar.
The one bragging appears to have deleted his post, but it was connected here to a brief but good thread :
https://x.com/DoloresGMorris/status/2058943409801371860
It doesn't seem to be something that most authors are pursuing. If you publish something interesting that gives others ideas for extending the research, you probably will get cited, but is that the motivation or advancing knowledge?
Because knowledge won't be advanced by hallucinating AI.