{"id":109427,"date":"2026-07-17T11:11:22","date_gmt":"2026-07-17T11:11:22","guid":{"rendered":"https:\/\/www.europesays.com\/ai\/109427\/"},"modified":"2026-07-17T11:11:22","modified_gmt":"2026-07-17T11:11:22","slug":"ai-made-up-a-science-term-now-its-in-dozens-of-papers","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ai\/109427\/","title":{"rendered":"AI Made Up a Science Term \u2014 Now It\u2019s in Dozens of Papers"},"content":{"rendered":"

\"AI-generated<\/a>AI \u201cdigital fossils\u201d are already polluting the world. They seem authentic and complex, but in fact, they don\u2019t make sense. Kind of like this AI-generated image.<\/p>\n

In 2023 and 2024, as AI text generators started to become mainstream, a curious trend emerged: the word \u201cdelve\u201d began appearing in a suspicious number of science papers. It became a kind of calling card for AI-generated content \u2014 but it\u2019s far from the weirdest one.<\/p>\n

Let us introduce you to: \u201cvegetative electron microscopy.\u201d<\/p>\n

Vegetative what?<\/p>\n

If you know basic science, you\u2019re already raising an eyebrow. \u201cVegetative electron microscopy\u201d doesn\u2019t make sense \u2014 and that\u2019s because it isn\u2019t a real thing. It\u2019s what researchers call a \u201cdigital fossil\u201d \u2014 a strange, erroneous term born from a mix of optical scanning errors and an apparent translation mistake, then preserved in the data used to train artificial intelligence.<\/p>\n

Remarkably, this nonsense phrase appears to have emerged independently through two unrelated errors.<\/p>\n

Back in the 1950s, two papers<\/a> in the journal Bacteriological Reviews were scanned and digitized. In one of them, the word \u201cvegetative\u201d appeared in one column and \u201celectron microscopy\u201d in the adjacent one. The OCR software mistakenly merged the two \u2014 and so, the fossil was born.<\/p>\n

\"\"<\/a><\/p>\n

Then, in 2017<\/a> and 2019<\/a>, two papers used the term again. Here, this appears to be a translation error<\/a>. In Farsi, the words for \u201cvegetative\u201d and \u201cscanning\u201d differ by only a single dot. So instead of scanning electron microscopy, you got vegetative electron microscopy.<\/p>\n

\"\"<\/a><\/p>\n

All of this came to light thanks to a detailed investigation by Retraction Watch in February. But this wasn\u2019t the end of the story.<\/p>\n

Why this matters<\/p>\n

You\u2019d think this weird glitch wouldn\u2019t matter \u2014 but it turns out, it kind of does.<\/p>\n

\u00d7<\/p>\n

Thank you! One more thing…<\/p>\n

Please check your inbox and confirm your subscription.<\/p>\n

The term has now appeared in at least 22 different papers<\/a>. Some have been corrected or retracted, but by then, the damage was done. Even El Pa\u00eds, one of Spain\u2019s leading newspapers, quoted it in a story<\/a> in 2023.<\/p>\n

Why? Blame AI.<\/p>\n

Modern AI systems are trained on vast troves of data \u2014 essentially everything<\/a> they can scrape. Once \u201cvegetative electron microscopy\u201d appeared in several published sources, the AI models treated it like a legitimate term. So when researchers asked these systems to help write or draft papers, the models sometimes spat it out, blissfully unaware that it was gibberish.<\/p>\n

According to Aaron J. Snoswell and colleagues, who published a deep dive on The Conversation<\/a>, the term began polluting the AI knowledge pool after 2020 \u2014 after those two problematic Farsi translations. And it\u2019s not just a one-time fluke: the error persists in large models like GPT-4o and Claude 3.5.<\/p>\n

\u201cWe also found the error persists in later models including GPT-4o and Anthropic\u2019s Claude 3.5,\u201d the group write<\/a> in a post on The Conversation. \u201cThis suggests the nonsense term may now be permanently embedded in AI knowledge bases.\u201d<\/p>\n

\u201cPermanently\u201d may be too strong in a literal sense. AI developers can filter outputs, revise training sets and replace old models. But correcting an error buried inside billions or trillions of learned statistical relationships is much harder than fixing a line in a conventional database. Removing the original phrase from the internet wouldn\u2019t necessarily remove the association from models already trained on it.<\/p>\n

AI-assisted scientific writing is no longer a marginal phenomenon<\/p>\n

This bizarre example is more than a fun anecdote \u2014 it highlights real risks.<\/p>\n

Since the vegetative-electron-microscopy story emerged, researchers have begun measuring AI\u2019s influence across entire bodies of scientific literature. <\/p>\n

In 2025, researchers analyzed<\/a> more than 15 million biomedical abstracts indexed in PubMed. They tracked vocabulary that became unusually common after the release of ChatGPT and estimated that at least 13.5 percent of abstracts published in 2024 had been processed with a large language model. Rates varied widely between fields, journals and countries, sometimes reaching about 40 percent.<\/p>\n

Another study, published in Nature Human Behaviour, examined more than<\/a> 1.1 million papers and preprints. By September 2024, the researchers estimated that large language models had modified 22.5 percent of computer-science abstract sentences and 19.6 percent of introduction sentences. The estimated rates were lower in mathematics and Nature Portfolio journals but had still risen sharply after ChatGPT\u2019s release.<\/p>\n

In 2026, an analysis of 7.3 million articles published between 2020 and 2025 found evidence of some degree of LLM influence in slightly more than half of the papers from 2025<\/a>. That did not mean that half of the papers were wholly written by chatbots. The category included everything from light linguistic editing to much more extensive generation.<\/p>\n

Using AI to improve grammar or phrasing isn\u2019t scientific misconduct. For many researchers, especially those writing in a second language, these tools can make papers clearer and more accessible. The problem arises when authors use them without checking the output, allow them to invent references or scientific claims, or conceal the extent to which a machine produced the paper\u2019s reasoning.<\/p>\n

An AI Arms Race<\/p>\n

Researchers are trying to fight this and detect this sort of issue. The Problematic Paper Screener<\/a>, for instance, is an automated tool that combs through 130 million articles<\/a> every week. It uses nine detectors searching for new instances of known fingerprints or improper use of AI. They found 78 papers<\/a> in Springer Nature\u2019s Environmental Science and Pollution Research alone.<\/p>\n

But it\u2019s an uphill battle.<\/p>\n

There\u2019s already so much AI content everywhere that it\u2019s almost becoming virtually impossible to detect it; and that\u2019s just one part of the problem. Scientific journals are another problem.<\/p>\n

Journals have every incentive to protect their reputation and avoid retractions, even if it means defending dubious content. Case in point: Elsevier initially tried to justify<\/a> the use of \u201cvegetative electron microscopy\u201d before ultimately issuing a correction. They ultimately issued a correction but the response is telling.<\/p>\n

The problem is that as long as tech companies aren\u2019t transparent about their training data and methods, researchers have to play detective and look for AI needles in the publishing haystack. According to one estimate<\/a>, there are close to 3 million papers published a year, and the use of AI in writing is becoming more and more common.<\/p>\n

The real danger is that these kinds of accidental errors can become entrenched in our scientific record \u2014 and once embedded, AI systems will keep repeating them. Knowledge is incremental, and if we build on wrong foundations, the consequences can be severe.<\/p>\n

Ultimately, it seems even nonsense, once digitized and published, can become immortal.<\/p>\n","protected":false},"excerpt":{"rendered":"AI \u201cdigital fossils\u201d are already polluting the world. They seem authentic and complex, but in fact, they don\u2019t…\n","protected":false},"author":2,"featured_media":109428,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[24,2720,32051,9439,25,56408,56409,11206,56410,56411,56412,56413,32902],"class_list":["post-109427","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ai","tag-ai","tag-ai-hallucinations","tag-ai-in-science","tag-ai-training-data","tag-artificial-intelligence","tag-claude-3-5","tag-digital-fossils","tag-gpt-4o","tag-knowledge-integrity","tag-ocr-errors","tag-paper-retractions","tag-retraction-watch","tag-scientific-publishing"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/109427","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=109427"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/109427\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/109428"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=109427"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=109427"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=109427"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}