Holy cow, this is bad - check out the nonsensical figures in this review: completely generated in Midjourney (!!), not even re-labeling the gibberish text labels (!!!).
This was published by Frontiers in Cell and Developmental Biology, where the editorial office should have caught this and nuked it from orbit. Wow.
https://www.frontiersin.org/articles/10.3389/fcell.2023.1339390/full
@moritz_negwer @academicchatter This has probably been peer reviewed by Dr. D. Alle and Dr. Chad Gepetee.
@HolgerTDittmann @academicchatter If only! Frontiers actually publishes the reviewer's names, in the hope that the reputational risk leads to stricter enforcement. Not so in this case.
Vice apparently interviewed one of the reviewers, and he essentially said "not my problem": https://www.vice.com/en/article/dy3jbz/scientific-journal-frontiers-publishes-ai-generated-rat-with-gigantic-penis-in-worrying-incident
@moritz_negwer @HolgerTDittmann @academicchatter
"I only review the paper based on its scientific aspects"
What does that even mean?
That "iollotte sserotogmar cell" and "dck" are all good scientific aspects?
If I was looking for papers in that field, I'd probably discard anything anything the authors or reviewers had a hand in. How could they be trusted?
@richlv @HolgerTDittmann @academicchatter I mean I agree, the reviewer messed up here. But the entire chain of supposed checks before them did too.
I would argue that this is an especially egregious example of a wider trend. The current peer review system is not well prepared for the coming onslaught of cheap low-effort GenAI sludge.
Combine cheap generation and insufficiently calibrated bullshit detectors, plus huge incentives for publication, and we'll keep seeing more like this.
@richlv @HolgerTDittmann @academicchatter Related: Adam Mastroianni argues in these blog posts that peer review is conceptually unfit to weed out bad science (and that may not matter all that much in the grand scheme of things). Well worth a read:
The rise and fall of peer review
https://www.experimental-history.com/p/the-rise-and-fall-of-peer-review
The dance of the naked emperors
https://www.experimental-history.com/p/the-dance-of-the-naked-emperors
Science is a strong-link problem
https://www.experimental-history.com/p/science-is-a-strong-link-problem
@moritz_negwer
Those are good reads. But I have a hunch Adam should have a look into mediocre-link, e.g. noise problems. I agree on lots of his findings, but to me he seems to underestimate the detrimental effect of noise (like this AI BS, but also lots of Social Media chatter) on public - or in this case scientific - discourse.
@richlv @HolgerTDittmann @academicchatter
