49
Scientists aghast at bizarre AI rat with huge genitals in peer-reviewed article (arstechnica.com)
submitted 10 months ago* (last edited 10 months ago) by blakestacey@awful.systems to c/techtakes@awful.systems
NSFW 28 comments fedilink hide all child comments

Appall and scorn ripped through scientists' social media networks Thursday as several egregiously bad AI-generated figures circulated from a peer-reviewed article recently published in a reputable journal. Those figures—which the authors acknowledge in the article's text were made by Midjourney—are all uninterpretable. They contain gibberish text and, most strikingly, one includes an image of a rat with grotesquely large and bizarre genitals, as well as a text label of "dck."

A dck pck, if you will.

Count me among the "some scientists online" who "questioned whether the text was also AI-generated". I mean, it's a disjointed mess. Right off, we get this:

The term “stem cell” was first coined in 1901 by Regaud

Um, no. But if that could be taken for human error, what about a sentence like this:

They were physically sheared and digested with a solution of DnaseI, hyaluronidase, collagenase, and trypsin using a two-step enzymatic digestion method in which the digestive enzymes included DnaseI, hyaluronidase, collagenase, and trypsin.

Just a bit before that, the text does a swerve into what sounds like a specific experiment, which doesn't fit with its surroundings and is very strange in a review article. My guess is the whole thing was made by stitching together LLM responses.

The publisher, Frontiers Media, is not exactly held in high regard overall.

all 30 comments
sorted by: hot top controversial new old
[-] bitofhope@awful.systems 30 points 10 months ago

Here's the pic in its full glory.

AI generated illustration of a rat with a massive schlong in cross-section and misplaced and misspelled captions: "sentollc stem cells", "Dissilced", "dislocctal stem ells" with the first letter s mirrored, "Rat", "Testtomcels", "dck", "Retat", "iollotte sserotgomar cell", "Spermatocial syem cells", and "Sterrm cells" with the last letters smeared together into some kind of quad-hump letter m

Tag yourself, I'm Testtomcels

[-] bitofhope@awful.systems 20 points 10 months ago

An AI illustration of something vaguely biotechnological with lots of incomprehensible labels. Looks a bit like a circuit board.

This one from the same is also funny

[-] self@awful.systems 19 points 10 months ago

all leftists care about is propronounization stat protemns, translocation, and posting on DMmer

[-] acausal_masochist@awful.systems 10 points 10 months ago

Reminds me of the plastic placemats I ate off of as a toddler.

[-] Syn_Attck@lemmy.today 2 points 8 months ago

Jak stats and prom, what a night.

[-] self@awful.systems 11 points 10 months ago

di>locttal stem ells was my father’s name. please, call me dck

like fuck can you imagine looking at nothing but results like these and believing “yep these are the glimmerings of AGI, what a revolutionary technology”

[-] fasterandworse@awful.systems 10 points 10 months ago

this will be in a future gallery exhibition about the beginning of the Art Cynique movement

[-] sailor_sega_saturn@awful.systems 8 points 10 months ago

I go by many names, but you may call me Iollotte.

[-] sc_griffith@awful.systems 7 points 10 months ago

bond voice I'm retat. rat retat.

[-] V0ldek@awful.systems 18 points 10 months ago

According to researchers referenced in a 2015 blog post quoted by Allison and James Kaufman in the 2018 book Pseudoscience: The Conspiracy Against Science, "Frontiers has used an in-house journals management software that does not give reviewers the option to recommend the rejection of manuscripts" and the "system is setup to make it almost impossible to reject papers". source

Ah yes, pear reviw.

[-] bitofhope@awful.systems 10 points 10 months ago

Recommend publication?
>yes
later

[-] blakestacey@awful.systems 13 points 10 months ago

And it's been retracted:

Following publication, concerns were raised regarding the nature of its AI-generated figures. The article does not meet the standards of editorial and scientific rigor for Frontiers in Cell and Development Biology; therefore, the article has been retracted.

This retraction was approved by the Chief Executive Editor of Frontiers. Frontiers would like to thank the concerned readers who contacted us regarding the published article.

[-] sailor_sega_saturn@awful.systems 13 points 10 months ago

At least they didn't claim to have "high standards", only "standards"!

[-] froztbyte@awful.systems 7 points 10 months ago

That wording is a work of art

[-] self@awful.systems 10 points 10 months ago

A dck pck, if you will.

I am so glad it wasn’t just my brain that went there

I read an article about this on mastodon earlier, but somehow it seems like it took a long time for anyone to read the text of the paper and realize that’s garbage too

[-] jonhendry@awful.systems 10 points 10 months ago

I'm not sure the "frontiers in" journals are all that reputable.

[-] blakestacey@awful.systems 9 points 10 months ago

My informal impression is that they range from "OK" to "... the Hell?!".

[-] jonhendry@awful.systems 9 points 10 months ago

To be fair though rodents can have pretty huge balls. Like dragging on the ground behind them huge.

[-] sc_griffith@awful.systems 9 points 10 months ago* (last edited 10 months ago)

same with our brave soldiers. god bless

[-] Communist@lemmy.ml 8 points 10 months ago

have you seen the photo because we're talking so big that the rat can't reach its little feet to the ground huge

[-] dgerard@awful.systems 9 points 10 months ago

Frontiers Media is the second entry of the "don't cite this BS" list for Wikipedia.

[-] swlabr@awful.systems 7 points 10 months ago

Funny, when I ask chatgpt to draw a rationalist, the same thing pops up.

[-] CommanderCloon@lemmy.ml 6 points 10 months ago

New ppb for Hexbear users?

[-] autotldr 1 points 10 months ago

This is the best summary I could come up with:

Appall and scorn ripped through scientists' social media networks Thursday as several egregiously bad AI-generated figures circulated from a peer-reviewed article recently published in a reputable journal.

But, looking closer only reveals more flaws, including the labels "dissilced," Stemm cells," "iollotte sserotgomar," and "dck."

Many researchers expressed surprise and dismay that such a blatantly bad AI-generated image could pass through the peer-review system and whatever internal processing is in place at the journal.

One scientific integrity expert questioned whether it provide an overly complicated explanation of "how to make a donut with colorful sprinkles."

The image is supposed to provide visual representations of how the signaling pathway from Figure 2 regulates the biological properties of spermatogonial stem cells.

As such, research journals have recently set new authorship guidelines for AI-generated text to try to address the problem.

The original article contains 496 words, the summary contains 137 words. Saved 72%. I'm a bot and I'm open source!

this post was submitted on 16 Feb 2024
49 points (100.0% liked)

TechTakes

1489 readers
30 users here now

Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.

This is not debate club. Unless it’s amusing debate.

For actually-good tech, you want our NotAwfulTech community

founded 2 years ago
MODERATORS
dgerard@awful.systems