Publishers Using AI-Written Abstracts of Your Work – CORRECTION
Academic publishing corporation Wiley appears to be using a large language model (such as ChatGPT) to generate abstracts for works it publishes, without involving the author—and they may not be good.
UPDATE (11/18/24): It turns out that, according to Wiley, AI was not used to author the abstract in question. According to Dr. Field, a representative of Wiley explained to her what happened:
These abstracts are not in fact AI-generated, but were generated by the content team prior to publication, which is very unusual and only done in the event that we are unable to receive abstracts written by contributors. Back in 2021 this happened for a couple of Companion projects (as soon as we were able to receive correct abstracts, we would immediately replace the faulty ones), and I’m not sure where in the process was the issue, and why contributors’ own abstracts weren’t successfully solicited, but it is standard for all such projects to have contributors’ own abstracts and keywords submitted with their chapters well before publication.
Wiley is reportedly expediting the process of replacing the abstracts.
Dr. Field writes, “I’m happy to say that my own proper abstract is now up on the Wiley-Blackwell Online Library.”
Sandra Leonie Field, a lecturer in philosophy at Monash University, reports that this happened to her piece, “Hobbes on Power and Gender Relations,” which appeared in A Companion to Hobbes.
The abstract that appears on the Wiley site for the chapter was not written or vetted by Field. Her own abstract for it is on Philpapers. Here they are:
Dr. Field writes:
That they did this without checking is annoying in itself. But it’s upsetting because their abstract is so bad, completely garbling my argument and making me sound like an idiot. In particular, the final sentence of the AI-abstract[*] is a quick gloss on Carole Pateman’s view; the AI-abstract gives the impression that is my final view, whereas in fact that is the view that I am critiquing. To add insult to injury, the AI-generated abstract genders me male (‘what he will call’, ‘his application of the model’)—in a paper talking about questions of gender.
Has this happened to you? Have you checked? Let us know if it has, and if so, which publishers are involved. Suggestions for addressing this issue are welcome.
*see above correction.
I’m just shocked that Wiley would do something like this.
I’m not shocked at all. All the big presses are getting in bed with generative AI.
I sense (perhaps mistakenly) sarcasm in SCM’s comment.
unbelievable
The irony being that people actually believed something that did not in fact happen. AI is becoming so good that we see it even where it’s not.
Don’t people who use these realize you should give everything it prints a second look!??
Worse: Letting an LLM write an abstract and then vetting it would be more work than just editing the author-written abstract that they already had— for no benefit whatsoever.
My students don’t seem to, despite my repeated warnings.
“We paid for the AI, and god-be-damned we’re going to use it.”
AI is a solution in search of a problem, so technology companies are currently forcibly bundling their AI products with their established software suites (and increasing the price as they go). I guess the hope is that their customers might find a use for it (since they themselves cannot).
To no-one’s surprise, we can strike “academic publishing” from the list. Thanks Wiley!
This is absolutely nuts…
Time for a tenured/senior prof boycott, friends. No more involvement with Wiley, no more submitting to publishers who do this stuff. This is just going to keep getting worse and after a while we will just look dumb for having played along for so long.
Not that, but I have received what I believe are AI “peer reviews.” Quick, if not terribly deep. What’s funny is that these “journals” ask for the names of a handful of scholars whom you think would be competent, but then send you “reviews” a couple days later — and it’s clear they didn’t have time even to request reviews from those people.
I mean, I ‘m not complaining. These are crappy journals, and I’ve no doubt that they will publish my papers. But it’s kind of funny receiving a positive “review 1” and a negative “review 2” precisely because that’s exactly what the editor requested of the its LLM.
Why do you want to publish in crappy journals?
Sensible question. They are older papers in areas I no longer study that I have largely given up on. I continue to try to get stuff I am currently working on published in decent journals. But with these older papers, it’s apparently either a crappy venue or a firepit–and I continue to think they might have at least SOME merit.
Maybe it’s silly…or just an old man’s hubris. But these places are free and quite a bit faster than reputable places. Publication is painless, and citation to them is possible (if unlikely). So…. what’s the harm?
If you are willing to publish in a journal that doesn’t really peer review stuff, and thus if you don’t really care about the imprimatur of serious peer review, why not just post your papers on PhilPapers.org? That is a way to disseminate them to probably a larger audience than via publishing in an unknown crummy journal, and it has the added advantage of not lending your reputation and intellectual effort to support what is effectively a fake journal.
SSRN, Medium, Philpaoers, “fake journals,” personal blogs….they are all vanity publications of one kind of another. I have used several. I suppose they all have different reaches.
But I won’t pay any of them, so I’m not sure how much “support” I’m providing.
Just wanted to add that I have been posting reviews of democracy books in Richard Marshall’s non-peer reviewed 3:16 AM Magazine for about 5 years now. I suppose, based on this scold, I should stop immediately. I will take it under advisement.
This seems like a false equivalence. Blogs and on-line publications have their place. The challenge was not to publishing somewhere without peer review, but rather to publishing in a venue that you yourself think of as “crappy”.
Well, crappy for a philosophy journal. But OK as one more distribution channel.