The pressure to publish or damage has led some troubled researchers to pay fake papers to fill out their CV.
Worse, some of these fake papers are being published in reputable scientific journals.
A computer program designed to detect this type of research shows that many continue to repeat what their peers are doing.
The study was published as a pre-printed paper and is awaiting peer review, but if the results are confirmed, they are very concerning.
Using artificial intelligence, researchers trained a computer to look for several red flags that are often seen in fraudulent papers submitted to scientific journals.
After the tool was able to pick out red flags with 90 percent accuracy, it was used to include nearly 5,000 neuroscience and medical papers published in 2020.
The tool listed 28 percent as manufactured or painted.
If this applied to all 1.3 million biomedical papers published in 2020, more than 300,000 would have been identified.
Not all of these flags are false, but they help identify more suspicious studies that should be looked at more closely by reviewers.
Out of every 100 papers with red flags for the new tool, about 63 were false, and 37 were true.
Psychologist Bernhard Sabel from the Otto-von-Guericke University of Magdeburg in Germany is one of the study’s lead authors and editor of the neurology journal.
He, like many others, has been struggling with the recent rise of fake papers. But even Sabel was surprised by the initial numbers of his instrument.
“It’s hard to believe,” he said Science.
Sabel and her colleagues are blaming the ‘paper makers’ for this fraud. Paper machines advertise themselves as ‘educational support’ services, but in reality, they use AI to develop and sell fake news to researchers.
The prices of counterfeit papers can range from US$1,000 to US$25,000.
The quality of these studies is often poor but sufficient to merit peer review, even in established journals.
Publishers know that this is a big issue that damages their reputation. Scientists have also persuaded the media to accept false papers to explain the problem.
In some cases, paper mills go so far as to pay publishers to accept their fake studies. In fact, this type of unsolicited email to a magazine editor led to a new study.
“Because the problem still seems to be small (about 1 in 10,000 publications), publishers and academic groups are starting to change the process, peer review, and publish,” the researchers write.
“Yet the extent of disinformation is unknown, even though the number of reports in the papers is increasing.”
Between 2010 and 2020, this new tool showed a 12 percent increase in the number of fake papers published by other journals.
The country with the most counterfeiters was China, which contributed to more than half of the red flags. Russia, Turkey, Egypt, and India also contributed significantly.
Researchers argue: “The publication of pseudoscience is the biggest fraud of all time, wasting money, delaying medical progress, and possibly putting lives at risk.”
And the rise of artificial intelligence like ChatGPT only makes fraud more dangerous.
In order to deal with this emerging technology and respect the history of science itself, researchers say a more comprehensive review process is urgently needed.
Preprint published in mdRxiv.
