Language Log

« previous post | next post »

This may sound hopelessly old-fashioned.  People were making the accusation more than half a century ago, but the same problems it points to persist even today.

In computer science, garbage in, garbage out (GIGO) is the concept that flawed, biased or poor quality ("garbage") information or input produces a result or output of similar ("garbage") quality. The saying points to the need to improve data quality in, for example, programming. Rubbish in, rubbish out (RIRO) is an alternate wording

The principle applies to all logical argumentation: soundness implies validity, but validity does not imply soundness. In essence, the logic or algorithm may be correct, but using flawed inputs (premises) is still an informal fallacy.

(WP)

The dangers of GIGO / RIRO have only been magnified with the advent of AI.

Garbage In, Garbage Out: Scientific Publishing Faces Pandemic-Like Spread of Fake Studies
Fake science is big business, and it’s spreading faster than the real thing.

HOLLIE McKAY
Mar 23, 2026, 12:02 PM ET
FacebookX (Twitter)EmailWhatsApp

————

A professor of engineering sciences and applied mathematics at Northwestern University, Luís Amaral, describes his latest research — on fraud in scientific publishing — as “probably the most depressing project I’ve been involved with in my entire life.” 

Despite having spent his career studying complex social systems — how institutions fracture and how incentives corrupt behavior at scale — Mr. Amaral says nothing prepared him for what he and his colleagues uncovered when they turned their analytical lens toward the global scientific publishing apparatus.

His recent study, published in the “Proceedings of the National Academy of Sciences,” found that organized scientific fraud, involving networks of paper mills, brokers, and compromised journals, is increasing rapidly and now outpaces the growth of legitimate publications. The repercussions of fraudulent science cannot be understated, he warns. 

“If we do not create awareness around this problem, worse and worse behavior will become normalized,” Mr. Amaral told reporters at his university. “At some point, it will be too late, and scientific literature will become completely poisoned. Some people worry that talking about this issue is attacking science. But I strongly believe we are defending science from bad actors.”

A Criminal Enterprise, Not a Rogue Lab

Peer-reviewed literature underpins medical treatments, public health guidance, pharmaceutical approvals, and policy decisions at every level of government. If the literature is being systematically poisoned, the downstream effects touch virtually everything that relies on the authority of science.

For decades, the public narrative around scientific fraud centered on the lone bad actor — the ambitious researcher who fudged a dataset, the graduate student who plagiarized a passage. That framing, the Northwestern team argues, is now dangerously outdated. 

For the co-founder of Retraction Watch, Ivan Oransky, the reckoning has been a long time coming.

“Publishers really need to acknowledge that they’ve known about paper mills since at least 2013,” Mr. Oransky told The New York Sun. “Now they’ve grown a lot, and they’ve industrialized. They don’t just sell papers. They sell authorships, citation manipulation, and ways to boost your standing in the rankings. And now, of course, they’re using AI to do even more of it.”

In their research, Mr. Amaral and his colleagues uncovered sophisticated global networks systematically undermining the integrity of academic publishing. At the center are paper mills, outfits functioning like production lines for academic manuscripts, selling papers to researchers who want to pad their publication records quickly. 

These manuscripts often contain fabricated data, manipulated or stolen images, plagiarized text, and sometimes claims that are scientifically impossible. Scientists can buy not just papers, but also citations — conjuring the appearance of a well-regarded academic career from nearly nothing.

By any measure, the market is profitable. Authorship slots sell for hundreds or even thousands of dollars. One survey of medical residents at tertiary hospitals in southwest China found that nearly 47 percent of respondents reported buying and selling papers or having others write them. Some major publishers have estimated that as many as one in seven submissions show signs of paper mill origin.

The incentives, Mr. Oransky explained, extend well beyond individual researchers, all the way up to the institutions employing them, creating a system in which fraud is not just tolerated but structurally rewarded.

“If you spend a few hundred, even a few thousand dollars to buy a paper that is guaranteed to be published, and then because of that you end up getting tenure or a promotion, as an investment, you can see how it makes sense,” he said. “Universities benefit too. The international university rankings are, I would argue, pernicious in their impact. It’s kind of like Lance Armstrong — it’s the same kind of doping, but in academia.”

The corruption runs deeper than publish-or-perish anxiety, said the chief editor of the “Radiation and Health” journal at Frontiers in Public Health and adjunct professor of biochemistry at the University of Helsinki, Professor Dariusz Leszczynski. 

“Financial incentives are a very important factor. However, not only is money the problem,” Mr. Leszczynski told the Sun. “Advancing careers through ‘beefing up’ publication records is pretty common. Also, climbing up the ladder and becoming a recognized expert through ‘fake citations’ is also common.”

The fraud does not stop at fabricated manuscripts. Brokers serve as intermediaries, connecting paper mills to compromised journals while bypassing quality controls. When legitimate journals prove too difficult to penetrate, these networks target defunct publications. 

“This happened to the journal ‘HIV Nursing,’” a postdoctoral researcher at Northwestern University, Reese Richardson, told the Sun. “It was formerly the journal of a professional nursing organization in the United Kingdom, but it stopped publishing, and its online domain lapsed. An organization bought the domain name and started publishing thousands of papers on subjects completely unrelated to nursing, all indexed in Scopus.”

Despite efforts to curb paper mills, suspected paper-milled articles are growing exponentially, doubling every one and a half years — compared to total publications, which double every 15 years. 

In 2023 alone, publishers retracted more than 10,000 papers, a record driven largely by the collapse of publisher Hindawi, which retracted over 8,000 articles after paper mills were found to have systematically infiltrated its journals, costing parent company Wiley an estimated $35 million to $40 million.

Selected readings

• "The wonders of Google Translate" (9/22/12)
• "Penglin Wang’s response to David Marjanović’s comments" (20/21/19)
• "Mandatory treatment for generic plurals?" (/13/09)
• "Ultraconserved words? Really??" (5/18/13)

[Thanks to John Rohsenow]

March 30, 2026 @ 3:38 am · Filed by Victor Mair under Artificial intelligence

Permalink