The recent audit of 2.5 million biomedical papers didn’t uncover a "surge" in fake citations. It uncovered the predictable, mathematical result of an industry that treats researchers like high-volume manufacturing plants. People are clutching their pearls over "paper mills" and "citation cartels" as if these are some foreign virus invading a pristine body of knowledge.
They aren't. They are the body’s response to a lethal incentive structure.
The standard narrative—pushed by legacy journals—is that we are facing a crisis of integrity. They want you to believe that "bad actors" are ruining science. That’s a convenient lie. It allows the gatekeepers to sell more detection tools and more "integrity services" while ignoring the fact that they built the hunger games that made these tactics necessary. If you measure a scholar's value by a metric that can be gamed, it will be gamed. Goodhart’s Law isn't a suggestion; it’s a death sentence for honest data.
The Myth of the "Innocent" Audit
When a major audit claims to find hundreds of thousands of fraudulent references, it frames the issue as a failure of policing. The logic goes: If we had better algorithms, we’d have cleaner science. Wrong.
I’ve watched departments burn through millions in grant funding chasing h-index targets while the actual "science" being produced was a secondary concern to the formatting of the bibliography. When the "publish or perish" culture moved from a metaphor to a literal career mandate, the market for fake citations didn't just appear—it was commissioned.
We don't have a "fake citation" problem. We have a measurement obsession problem.
In the biomedical field, citations are the currency of survival. They dictate who gets the lab, who gets the tenure, and who gets the NIH funding. By treating a citation as a "vote" for quality, we’ve effectively turned the scientific record into a high-stakes popularity contest where the voters can be bought for the price of a mid-range laptop and a VPN.
Why the "Fix" is Worse Than the Fraud
The proposed solutions are always more surveillance. More AI-driven screening. More bureaucratic layers. These "solutions" only widen the gap between wealthy institutions and the rest of the world.
Imagine a scenario where every paper must pass through a $5,000 "integrity audit" before publication. The paper mills won't stop; they’ll just get better at mimicking the patterns the AI is looking for. Meanwhile, the independent researcher or the underfunded lab in a developing nation gets priced out of the conversation.
The current "integrity" movement is becoming a protection racket for the status quo.
The Paper Mill is a Mirror
Everyone loves to hate on paper mills—those shadowy entities that churn out fabricated studies with purchased citations. But look at what they actually do. They produce exactly what the system asks for: high-volume, peer-reviewed (nominally), heavily-cited biomedical content.
If a fake paper can pass through the review process of a "reputable" journal and rack up fifty citations in two years, the problem isn't the mill. The problem is that the "reputable" journal is providing a product that is indistinguishable from fraud.
We are obsessed with the vessel of knowledge rather than the knowledge itself.
- Quantity over Clarity: A researcher with two groundbreaking papers is often viewed as less "productive" than one with fifty mediocre, self-referencing blurbs.
- Metric Manipulation: We use impact factors to judge journals, which leads journals to pressure authors to cite their previous work, creating an "official" version of the very citation cartels they claim to despise.
- The Peer Review Fallacy: We pretend that two overworked, unpaid volunteers reading a PDF for twenty minutes is a "gold standard" for truth.
Stop Trying to "Clean Up" Science
The "People Also Ask" sections of the web are full of queries like "How can I tell if a citation is fake?" or "What is the most reliable medical journal?"
These are the wrong questions. They assume that if we just filter out the "fake" stuff, the remaining "real" stuff will be pure. It won't. The "real" stuff is often just as biased, just as p-hacked, and just as influenced by funding as the blatant fabrications.
The only way to actually fix this is to burn the metric to the ground.
- Abolish the H-Index: Stop using a single number to decide if a human being is a good scientist. It is a reductive, lazy tool for administrators who don't want to actually read the work they are evaluating.
- Decouple Publication from Funding: When grants are awarded based on previous publication counts, you are incentivizing fraud. Period.
- End the Citation Fetish: A citation should be a map to a source, not a gold star. If we stopped counting them as "points," the market for fake ones would evaporate overnight.
The Brutal Reality of the Biomedical Record
Here is the truth nobody wants to say: A significant portion of the "legitimate" biomedical literature is already unusable.
Between the replication crisis and the pressure to produce "positive results," the delta between a "real" paper and a "paper mill" product is shrinking every day. The audit of 2.5 million papers didn't reveal a new threat; it revealed the logical conclusion of the last thirty years of academic policy.
The system isn't being attacked. It is fulfilling its design. If you want better science, stop measuring the noise and start looking at the signal. But that would require actually reading the papers, and apparently, nobody has time for that anymore.
Stop looking for "fake" citations and start asking why we made "real" ones so profitable to forge.
The audit is a distraction. The data is a symptom. The "surge" is just the sound of the machine running exactly as you built it.
