Typically, a scientific paper titled “Cellular functions of spermatogonial stem cells related to the JAK/STAT signaling pathway” does not attract a large readership at best. But if the article in question clearly features AI-generated images of him, and one of those images features a rat with frankly shockingly large genitals, well, well… That should be enough.
The paper was written by three Chinese researchers, co-authored with Dingjun Hao from Xi’am Hongui Hospital, and published online in the journal Frontiers in Cell and Developmental Biology (via Ars Technica). However, those who looked closely at the work after its release were stunned to discover an image of a rat with a penis and testicles so large that even the rat himself seemed to be looking at it in amazement.
Although this unfortunate rodent certainly attracted people’s attention, imagination, and scorn, twitter, wasn’t the only obvious sign that the AI image had inexplicably turned into a supposedly peer-reviewed work. The rat penis itself is labeled “cut off,” but my spell check and a quick online dictionary reference confirm that this is not a word. Meanwhile, another image is helpfully labeled “dck” but fortunately does not appear to feature a penis.
Other AI mislabeling issues include a misspelling of “stem cells” and a label pointing to the rat’s back labeled “Testtomcels.” I had to search to make sure this wasn’t an actual scientific term either. Again, I’m not the reviewer of this article, so I think it’s off the mark. Those concerned may have needed to be checked a little more carefully.
Ah, now I’m going to show you the image. Those of a more sensitive nature may want to avoid the following tweets.
I cannot stress this enough: unregulated GenAI will not only hinder, but harm, future progress in the arts as well as academia. Researchers published images using Generative AI. It was published in a peer-reviewed scientific journal. pic.twitter.com/bb0PDCf8iwFebruary 15, 2024
Well, we warned you.The article in question was immediately slammed on social media, saying it was clearly AI-generated and no human would do that…and it was some humans who actually trashed it. maybe Although he depicts clearly grotesque things, it is by no means an academic work.
The real concern here is how can something so blatantly false and frankly inappropriate get past a peer review process that many have long held sacrosanct? That’s true. This is said to be a process in which academic papers are checked and scrutinized by a team of experts in the same field to verify, to the best of their ability, that the paper is of high quality and to prepare it for publication.
If something as obvious as a rat with a really big problem can get by without attracting the attention of colleagues, then less obvious AI-generated images can go through that process and be incorporated into scholarly works. It seems possible. Much more likely.
Given that AI image generation is currently a mainstream activity and many people are experimenting with the tools currently available, if more sophisticated AI images slip past these supposedly rigorous checks, other science Of course, it may be used for research. As a result, the wormhole is completely opened up regarding the validity of data that has passed through the academic process and can potentially be cited, referenced, and become part of the current state of affairs.
For now, we can laugh at the poor surprised AI rat. But it is certainly sobering to think about the potential impact on the scientific community and the potential for inaccurate AI images to spread in important scientific research.
