[ad_1]
Horror and contempt swept scientists’ social media on Thursday as several egregiously bad AI-generated figures circulated from a peer reviewed article recently published in a renowned magazine. These figures, which the authors acknowledge in the text of the article 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 strange genitals, as well as a text label that says “dck.”
On Thursday, the editor of the review article, Frontiers, published an “expression of concern”, noting that he is aware of the concerns regarding the published article. “An investigation is currently underway and this notice will be updated accordingly once the investigation is concluded,” the editor wrote.
The paper in question is titled “Cellular functions of spermatogonial stem cells in relation to the JAK/STAT signaling pathway,” and was written by three researchers in China, including corresponding author Dingjun Hao of Xi’an Honghui Hospital. It was published online Tuesday in the journal Frontiers in Cell and Developmental Biology.
Frontiers did not immediately respond to Ars’ request for comment, but we will update this post with any response.
The first figure in the paper, the one containing the rat, immediately attracted attention when scientists began sharing it widely and commenting on it on social media platforms, including Bluesky and the platform formerly known as Twitter. From a distance, the anatomical picture is clearly wrong. But looking closer only reveals more flaws, including the tags “dissilced,” Stemm cell,” “iollotte sserotgomar,” and “dck.” Many researchers expressed surprise and dismay that such a blatantly bad AI-generated image could slip through. of the peer-review system and any internal processing that exists in the journal.
But the rat pack is far from the only problem. Figure 2 is less graphic but equally mangled. While intended to be a diagram of a complex signaling pathway, it is instead a mess. An expert in scientific integrity questioned whether it provides an overly complicated explanation of “how to make a sprinkle donut.” Like the first image, the diagram is littered with meaningless text and baffling images. Figure 3 is no better, offering a collage of small circular images densely annotated with gibberish. The image is supposed to provide visual representations of how the signaling pathway in Figure 2 regulates the biological properties of spermatogonial stem cells.
Some scientists online questioned whether the text was also generated by AI. One user noted that the AI detection software determined that it was likely to be generated by AI; However, as Ars previously reported, such software is unreliable.
The images, while egregious examples, highlight a growing problem in scientific publishing. A scientist’s success largely depends on his or her publication record, with a high volume of publications, frequent publications, and articles appearing in top-tier journals, all of which give scientists more prestige. The system incentivizes unscrupulous researchers to publish low-quality articles that, in the age of AI chatbots, could be generated with the help of AI. Researchers worry that the increasing use of AI will make published research less reliable. As such, research journals have recently established new authorship guidelines for AI-generated text to try to address the issue. But for now, as the Frontiers article shows, there are clearly some gaps.