Updated April 7, 2026
- Scientific fraud has become an industry, setting a 10,000-paper retraction record in 2023.
- AI tools let paper mills flood peer reviewers with submissions at unprecedented speed.
- "Publish or perish" culture gives researchers structural incentives to buy fraudulent publications.
Owen Brierley had seen this pattern before. As a senior lecturer at Kingston University studying digital systems and game engines, he recognized what happens when production scales faster than quality control. Now he was watching the same dynamics play out in academic publishing.
Scientific fraud, Brierley observed, has become an industry. Paper mills churn out formulaic research articles. Brokerages guarantee publication for a fee. Predatory journals bypass peer review entirely.
These operations disguise themselves as "editing services" or "academic consultants." Their business model depends on corrupting the scientific process.
Key figure
10,000+
Research papers retracted globally in 2023, setting a new record
The Numbers Tell the Story
The scale is remarkable. In 2023, global retractions exceeded 10,000 papers for the first time, according to Retraction Watch. Batch retractions–where publishers withdraw multiple papers at once–have surged since the 1990s, when they were rare. The trend reflects how fraud has become industrialized.
Publisher Hindawi alone retracted more than 8,000 articles in 2023, more than all publishers combined had retracted in any previous year. The cleanup cost its parent company Wiley an estimated $35-40 million in lost revenue.
What is a paper mill?
Paper mills are businesses that produce fake or low-quality research articles for sale. They operate like content farms, flooding journals with submissions. Some practice "journal targeting" (sending multiple papers to one publication) and "journal hopping" (submitting the same paper to multiple outlets). If even a fraction slip through peer review, the mill profits.
AI Accelerates the Arms Race
Generative AI has supercharged the problem. Researchers are witnessing an explosion of papers that exploit AI software to mine public datasets and generate text at unprecedented speed.
These hastily produced articles bear hallmarks of paper mill production: evidence fabrication, data manipulation, ethics misconduct, outright plagiarism.
Where peer reviewers once received ten submissions annually for a conference or journal, they now face 30 to 40 in six months or less. Legitimate research gets buried in the avalanche. Overwhelmed reviewers are themselves tempted to use AI tools to summarize papers and write reviews. The result is an arms race.
Science reported that paper mills have been bribing journal editors, with documented payments reaching $20,000, to ensure their clients' work gets published. Investigations found more than 30 editors at reputable journals who were complicit.
The paper mills industry is causing harm to scientific publishing. Read also our article on how the quality controls for scientific papers are breaking down. (Science Reader)
The Systemic Roots
Blaming individual bad actors misses the point, Brierley argues. The "publish or perish" culture creates structural incentives for fraud. Researchers need publications to win funding but need funding to conduct publishable research. Governments facing financial struggles cut research budgets, intensifying competition.
Related reading
AI Materials Discovery: 5 Things to Know
The gap between what AI predicts and what chemists can actually make has become the central drama of materials science.
→Globalization means individual researchers are lost in an ocean of competing voices. The temptation to game the system grows stronger. In this environment, the promise of guaranteed publication can seem like a lifeline.
The Retraction Watch Database now contains more than 55,000 entries. Studies suggest up to 2% of all publications in 2022 came from paper mills. Genuine mistakes in the scientific process accounted for only about 6% of retractions over the last decade.
If fraudulent publications continue growing faster than legitimate science, and AI-generated content keeps overwhelming human review capacity, the network of checks and balances that makes science work may not hold.
The question is whether reforms to publication metrics, funding mechanisms, and career incentives can arrive before the system breaks.
The Geography of Fraud
Paper mills thrive where governments reward volume over substance. The pattern is consistent across continents: tie career advancement to publication counts, and fraud fills the gap between what researchers can produce and what the system demands.
China offers the starkest example. Until 2020, Chinese universities paid cash bonuses for publications in prestigious journals. A single paper in Nature or Science could earn a researcher up to $165,000, nearly twenty times the average professor's annual salary. The incentive was blunt, and the response predictable. A cross-sectional study of cancer research literature found that 35.7% of flagged paper mill publications traced back to Chinese first authors. Beijing banned the bonuses, but the infrastructure they built persists.
Russia followed a similar script. Project 5-100, launched in 2013, aimed to place five Russian universities in the global top 100 by 2020. It tied funding to publication metrics. Universities that participated showed significantly higher retraction rates than those that did not. The program created what one analysis called a circular business model: paper mills collect fees from scholars, who collect bonuses from universities, which collect budget transfers from the state.
Iran accounts for roughly 20% of flagged publications in the same cancer literature study. Saudi Arabia, Egypt, and Pakistan each exceed 13%. The common thread is not geography or culture. It is policy. Wherever metrics replace judgment, mills find customers.
What Happened to Hindawi
The collapse of Hindawi is the clearest case study of what paper mills can do to a publisher. Wiley acquired the open-access giant for $298 million in 2021, expecting a portfolio of 200 journals to anchor its growth strategy. Within two years, that investment was in ruins.
The first signs appeared in the fall of 2022, when Hindawi retracted more than 500 articles for manipulated peer review involving compromised editors. By May 2023, Wiley had shuttered four journals to stop further contamination. The retractions kept climbing. By the end of 2023, the count exceeded 8,000 articles, more than all publishers combined had retracted in any previous year.
The financial damage was severe. Wiley reported $103 million in impairment charges during the first quarter of fiscal 2024 and an $18 million decline in research revenue attributed to the Hindawi portfolio. CEO Brian Napack, who had overseen the acquisition, stepped down in October 2023.
In December 2023, Wiley killed the Hindawi brand entirely. Five months later, it closed 19 more journals. The total retraction count reached 11,300. For the open-access publishing model, which depends on author trust and institutional credibility, Hindawi became a cautionary tale about what happens when quality control falls behind production speed.
Detection Tools and the New Arms Race
The counteroffensive is building, though it remains outmatched. The STM Integrity Hub, a consortium of 40 publishers, now screens more than 125,000 manuscripts per month and intercepts roughly 1,000 suspected paper mill submissions in each cycle. Wiley alone rejects 600 to 1,000 manuscripts per month before peer review, using a six-tool screening system that flags up to one in seven submissions across hundreds of its journals.
Some of the most effective detection work has come from individual researchers. Guillaume Cabanac, a computer scientist at the University of Toulouse, built the Problematic Paper Screener, which combs 130 million publications weekly for "tortured phrases," the telltale nonsense that emerges when paper mills run stolen text through paraphrasing software. "Kidney disappointment" for kidney failure. "Counterfeit neural organizations" for artificial neural networks. The tool has flagged more than 15,000 papers. Only 2,760 have been retracted.
Generative AI complicates this progress. Large language models produce fluent prose without the clumsy synonymization that made earlier paper mill output detectable. Roughly one in five computer science papers now shows signs of LLM involvement, and detection tools struggle with text that has been paraphrased or "humanized" by secondary AI tools.
The screening systems catch what they can. The question is whether detection can scale as fast as the fraud.
Sources
- Primary Source: How bad science is becoming big business (The Conversation, Owen Brierley)
- Additional Context:
- Science Magazine: Paper Mills Article by Retraction Watch Editor (Science)
- Hindawi reveals process for retracting more than 8,000 paper mill articles (Retraction Watch)
- The 'publish or perish' mentality is fuelling research paper retractions (The Conversation)
- The Truth about China's Cash-for-Publication Policy (MIT Technology Review)
- Retracted papers originating from paper mills: cross sectional study (BMJ/PMC)
- Scientometric indicators and the Russian university excellence initiative (Scientometrics)
- Brian Napack Out as Wiley CEO (Publishers Weekly)
- Wiley shuts 19 scholarly journals amid AI paper mill problem (The Register)
- STM Integrity Hub (STM Association)
- Up to one in seven submissions flagged by new paper mill tool (Retraction Watch)
- Problematic Paper Screener: Trawling for fraud (The Conversation)
- One-fifth of computer science papers may include AI content (Science)
Fact Check: Claim-by-Claim Verification Verified
The recap accurately represents the primary source article and supporting claims from reliable sources on retractions, paper mills, and related issues without factual errors.
Commentary
- The 2% paper mill estimate for 2022 publications aligns with studies but represents an approximation, not exact data.
- Review workload increases (10 to 30-40 submissions) reflect observed trends but are anecdotal from the source.
- Owen Brierley's affiliation with Kingston University and his observations on "publish or perish" are correctly attributed.
Sources used for verification
Academic/Peer-reviewed:
- How bad science is becoming big business - theconversation.com
Other reliable sources:
- Hindawi reveals process for retracting more than 8,000 paper mill articles - retractionwatch.com
- Paper mills are bribing editors at scholarly journals - science.org
- Retraction Watch Database - retractionwatch.com
Fact-checked by Perplexity Sonar Pro on 2026-01-23
