Paper Mills: An Old Crisis in Academia Made New
A recent article by Jeffrey Brainard at Science Magazine has put the spotlight on both the prominence and the role of paper mills in academic publishing.
The article highlights the work of neuropsychologist Bernhard Sabel, who screened some 5,000 neuroscience papers, finding that 34% of them were likely either made up or plagiarized from other sources. In the field of medicine, that number was 24%.
The culprit for this massive percentage of questionable papers is paper mills, companies that generate or plagiarize for academics who want or feel the need to publish papers, but don’t want to actually do the research that goes into them.
By Sabel’s admission, his tool is simplistic and is prone to a high false positive rate. However, even if half the papers are found to be false positives, that still leaves the number many times higher than the 2% estimated for most journals by the Committee on Publication Ethics (COPE) and the International Association of Scientific, Technical, and Medical Publishers (STM).
However, while that is what the report says, COPE and STM’s message was actually more nuanced. It said, “Most journals will see 2% suspected fake papers submitted and then for journals where paper mills have been successful in getting papers accepted, they see a sharp increase in suspect submissions”
According to that report, the highest percentage is actually 46%, with an average percentage of 14%.
However, that still puts the 34% found by Brainard in a shocking light, far higher than any estimate for an average rate.
In response to this issue, the academic publishing industry is working to fight back. Twenty publishers, including Elsevier, Springer Nature, and Wiley, are working together to develop Integrity Hub, a set of tools that, among other things, is expected to help weed out essay mill papers.
Likewise, STM, which represents 120 publishers, is preparing to launch its own tool that it hopes will detect manuscripts that were sent to multiple journals, an unethical practice that is often a warning sign of a paper mill work.
But while the publishing industry is finding ways to fight back, it’s facing a foe that is both very old and very young at the same time. It’s a nimble industry that has proved, if anything, very adaptable.
Making an Old Problem New
To be clear, the problem of paper mills in academic publishing is nothing new. I covered the topic in depth in March 2021, and highlighted that it was an issue that was being tracked as far back as 2007.
Simply put, paper mills exist in academic publishing for the exact same reason that essay mills exist in the classroom: To enable “authors” to gain the benefits of a paper without requiring them to do the work.
Both have had similar arcs. They increased rapidly as the internet grew and improved the ease of access to such mills and made it easier for would-be plagiarists to connect with inexpensive labor all over the world. Likewise, they both face existential challenges from AI, which may provide easier and cheaper ways for unethical authors to bypass the system.
However, one area where the paper mill industry has stood out has been its adaptability.
The Science Magazine article also highlighted the research of Adam Day, who owns Clear Skies, a company that offers a service named Papermill Alarm.
According to a blog post by Day, paper mills are incredibly quick to adapt and pivot. For example, when one journal was added to a Chinese Academy of Sciences (CAS) list of questionable journals, paper mill publications to the journal dropped to almost zero within months.
Another journal began to see a marked increase in paper mill submissions in 2021 and 2022, only to drop to almost nothing mere weeks after it was added to a CAS list.
Ironically, according to Day, the CAS lists are probably not reliable as journals featured in it likely have the fewer paper mill papers, simply because the mills have moved on.
Day does note that he does not know the criteria for the CAS and if paper mills are part of the process. That said, the impact those lists have is very clear.
But this puts academic publishing in a difficult position: How do you stop paper mills?
Fixing the Issue
Academic paper mills have a great in common with internet spammers. Since their product requires almost no work, they’re free to send out large quantities of garbage content, knowing that nearly all of it will be rejected or ignored.
Then, they simply make note of what did work and then repeat those steps until they stop working and then find new approaches.
What this means is that, even if the publishing industry manages to be 99% effective at detecting and stopping paper mill works, 1% will get through, and the mills will focus future efforts around that.
As several have described it, it’s a game of cat and mouse. Technological solutions can and should be used to mitigate the problem, but it’s not enough to actually stop it.
What has to change is the “publish or perish” environment many in academia work under. Such an environment puts the focus on quantity, not quality, of publications.
To that end, paper mills are just one of the symptoms of the problem. Predatory journals routinely dupe researchers into paying high open access fees for little benefit, researchers add their names to papers they had little to do with (sometimes for a joke) and researchers will break apart one paper into multiple smaller ones, taking up valuable publication space.
Without a change on the incentive side of research, there will be no change in all these practices. When researchers are rewarded for the quantity of their work, the quality inevitably suffers as the focus is racking up big numbers, not doing meaningful work.
Bottom Line
To be clear, there’s no simple answer here. While it’s easy for someone like me to say, “We have to fix the publish-or-perish environment,” doing so is incredibly difficult.
Gauging the quality of research is difficult and requires significant time from experts in that field. It’s much easier to look at a tally of publications and call that a “good enough” estimation of output. This is especially true for schools, where resources are often very limited.
However, it’s important to always be mindful of what our incentives are rewarding and how those incentive systems can be gamed. The current system rewards a variety of unethical behaviors, including paper mills.
If it weren’t for these incentives, paper mills wouldn’t be the multi-million dollar industry that it is today. It wouldn’t be a problem that potentially impacts 34% of neuroscience papers. It would be nothing more than an outlier, a problem that rarely comes up and is rarely discussed.
To that end, if paper mills do disappear, it won’t be because publishers beat them, it will be because AI replaced them. Simply put, there will always be journals that will publish mill papers and the paper mills will find them, through trial and error if needed.
Simply put, paper mills have been around for decades and have proven themselves resourceful and quick to pivot. That isn’t going to change any time soon.
The only thing that will change is what weaknesses they exploit…
Want to Reuse or Republish this Content?
If you want to feature this article in your site, classroom or elsewhere, just let us know! We usually grant permission within 24 hours.