It is almost impossible to overestimate the importance and impact of arXiv, the scientific repository that, for a time, almost single-handedly justified the existence of the Internet. ArXiv (pronounced “archive” or “Arr-ex-eye-vee” depending on who you ask) is a repository of preprints, where, since 1991, scientists and researchers have announced “hey, I just wrote this” to the rest of the scientific world. Peer review is moving slowly, but it is necessary. ArXiv simply requires a quick review from a moderator instead of scrutiny, so it adds an easy intermediate step between discovery and peer review, where all the latest discoveries and innovations can – with caution – be treated with the urgency they deserve more or less instantly.
But the use of AI has hurt ArXiv and it’s bleeding. And it is not certain that the bleeding can ever be stopped.
Inasmuch as recent article in The Atlantic Notes, Paul Ginsparg, creator of ArXiv and professor of information science at Cornell, has been concerned since the advent of ChatGPT that AI could be used to break down the small but necessary barriers preventing spam from being posted on ArXiv. Last year, Ginsparg collaborated on an analysis looking at likely AI in arXiv submissions. Horribly enough, scientists who were obviously using LLMs to generate seemingly plausible papers were more prolific than those who were not using AI. The number of articles from posters of work written or augmented by AI was 33% higher.
According to the analysis, AI can be legitimately used to, for example, overcome the language barrier. He continues:
“However, traditional signals of scientific quality, such as language complexity, are becoming unreliable indicators of merit, just as we are seeing an increase in the quantity of scientific work. As AI systems advance, they will challenge our fundamental assumptions about the quality of research, scientific communication, and the nature of intellectual work.”
It’s not just ArXiv. This is an overall difficult time for the reliability of scholarships in general. An astonishing self-property published last week in Nature described the misadventure of a clumsy scientist working in Germany, Marcel Bucher, who used ChatGPT to generate emails, course information, lectures and tests. As if that wasn’t enough, ChatGPT also helped him analyze student responses and was integrated into the interactive parts of his teaching. Then one day, Bucher tried to “temporarily” disable what he called the “data consent” option, and when ChatGPT suddenly deleted all the information it stored exclusively in the application, that is, on OpenAI’s servers, he complained in the pages of Nature that “two years of carefully structured academic work disappeared.”
Widespread AI-induced laziness, displayed in the exact area where rigor and attention to detail are expected and assumed, is a source of despair. We can assume that there is a problem when the number of publications rose just a few months after ChatGPT was first releasedbut now, as The Atlantic points out, we’re starting to get details about the real substance and scale of this problem — not so much Bucher-like, AI-wielding individuals who suffer from publication anxiety or perish and rush to publish a fake journal, but fraud on an industrial scale.
For example, in cancer research, bad actors can push for the publication of dull papers purporting to document “interactions between a tumor cell and a single protein among thousands that exist,” notes the Atlantic. If the article claims to be groundbreaking, that will raise eyebrows, which means the trick is more likely to be noticed, but if the bogus conclusion of the fake cancer experiment is ho-hum, that crap is much more likely to be published, even in a credible publication. All the better if it’s accompanied by AI-generated images of gel electrophoresis blobs that are also boring, but add extra plausibility at first glance.
In short, a flood of garbage has arrived in science, and everyone needs to become less lazy, from academics busy planning their courses to peer reviewers and ArXiv moderators. Otherwise, the knowledge repositories that once constituted one of the few reliable sources of information are about to be overwhelmed by the disease that has already – perhaps irrevocably – infected them. And does 2026 seem like a time when everyone, everywhere, becomes less lazy?
