The paper in question, “Cellular functions of spermatogonial stem cells in relation to JAK/STAT signaling pathway,” has seen a cacophony of reactions across scientific social networks as researchers shared their disbelief at the glaring oversight. The egregious flaws of the figures include labels such as “dissilced,” “Stemm cells,” “iollotte sserotgomar,” and, most shockingly, an explicit term for male genitalia. One figure included an image of a rat with genitals of disproportionate size, clearly illustrating the result of a flawed AI generation process.

Frontiers has since posted an “expression of concern” on their website and launched an investigation into the matter. The journal’s policies for authors do permit the use of generative AI but stipulate that any content created by the technology must be fact-checked for accuracy, which the authors acknowledged but evidently failed to adhere to in this instance.

The lapse has shed light on the broader issue of AI’s burgeoning role in scientific literature and the potential threat it poses to the integrity of academic publishing. It raises questions about the peer review process and the internal checks of journals that are supposed to prevent such instances. U.S.-based reviewer Jingbo Dai pointed out that his role was to review the paper’s scientific aspects, deferring the responsibility of vetting AI-generated figures to the publisher.

Elisabeth Bik, a science integrity consultant, weighed in on the situation, expressing concern that “if such botched illustrations can pass peer review so easily, more realistic-looking AI-generated figures have likely already infiltrated the scientific literature.” This incident, as Bik suggests, might just be the tip of the iceberg, indicating that the trustworthiness and value of scientific papers are at risk as AI technology continues to advance.

AI’s encroachment into academia is not new, with various journals updating their standards to accommodate the technology’s use. However, this incident underscores the need for more robust mechanisms to ensure that AI aids, rather than undermines, scientific discourse. As AI-generated text and images become more sophisticated, distinguishing between authentic research and fabricated content may become increasingly challenging, necessitating a re-evaluation of the publication and peer-review protocols.

Relevant articles: – Scientists aghast at bizarre AI rat with huge genitals in peer-reviewed article | It’s unclear how such egregiously bad images made it through peer-review.– Scientific Journal Publishes AI-Generated Rat with Gigantic Penis In Worrying Incident, VICE, Thu, 15 Feb 2024 20:40:08 GMT– ‘Rat Dck’ Among Gibberish AI Images Published in Science Journal, Gizmodo, Thu, 15 Feb 2024 20:47:00 GMT– The Impact of AI-Generated Imagery on Scientific Publications, ai2.news, Fri, 16 Feb 2024 02:12:40 GMT