Fake Science Proves Problematic for Academic Publishing as Journals Close From Research Fraud
Fake scientific papers, which are increasingly the product of paper mills using AI, undermine the credibility of all research.
I noted that City Journal did a detailed review of the ideological capture of science journalism. The must-read article ended by begging for a return to “the core principles of science—and the broader tradition of fact-based discourse and debate—our society.”
Facts are necessary for good science, the foundation of sensible and effective policy-making. Therefore, it is extremely troubling to learn that several academic journals were recently shuttered because their articles contained fake science.
Fake studies have flooded the publishers of top scientific journals, leading to thousands of retractions and millions of dollars in lost revenue. The biggest hit has come to Wiley, a 217-year-old publisher based in Hoboken, N.J., which Tuesday will announce that it is closing 19 journals, some of which were infected by large-scale research fraud.
In the past two years, Wiley has retracted more than 11,300 papers that appeared compromised, according to a spokesperson, and closed four journals. It isn’t alone: At least two other publishers have retracted hundreds of suspect papers each. Several others have pulled smaller clusters of bad papers.
Although this large-scale fraud represents a small percentage of submissions to journals, it threatens the legitimacy of the nearly $30 billion academic publishing industry and the credibility of science as a whole.
The discovery of nearly 900 fraudulent papers in 2022 at IOP Publishing, a physical sciences publisher, was a turning point for the nonprofit. “That really crystallized for us, everybody internally, everybody involved with the business,” said Kim Eggleton, head of peer review and research integrity at the publisher. “This is a real threat.”
The journals are associated with Wiley’s Hindawi subsidiary, an Egypt-based house that published 250 journals and was acquired in 2021. One contributing factor to the increasing number of fraudulent research papers is the use of artificial intelligence (AI).
Hindawi’s journals were found to be publishing papers from paper mills – organizations or groups of individuals who try to subvert the academic publishing process for financial gain. Over the past two years, a Wiley spokesperson told The Register, the publisher has retracted more than 11,300 papers from its Hindawi portfolio.
As described in a Wiley-authored white paper published last December, “Tackling publication manipulation at scale: Hindawi’s journey and lessons for academic publishing,” paper mills rely on various unethical practices – such as the use of AI in manuscript fabrication and image manipulations, and gaming the peer review process.
…The increasing availability and sophistication of generative AI is not the only factor contributing to the academic publishing crisis, but AI tools make fakery easier.
“The industry recognizes that AI is utilized by paper mills to generate fraudulent content,” Wiley’s spokesperson told us. “We’ve recently introduced a new screening technology that helps identify papers with potential misuse of generative AI before the point of publication.”
Professional organizations are asking if AI can be used to detect AI-generated paper-milled articles.
The answer appears to be not completely.
Using artificial intelligence, researchers trained a computer to look for several red flags commonly seen in fake papers submitted to scientific journals.
When the tool could pick out red flags with 90 percent accuracy, it was used to comb through roughly 5,000 neuroscience and medical papers published in 2020.
The tool marked 28 percent as probably made-up or plagiarized.
A second, hands-on check was, of course, required — and the second check found that two-thirds of the computer-flagged papers were indeed fake. So the computer can point out a basis for investigation but it can’t do the whole job.
Paper mills use AI to produce fake publications, which are then sold to researchers. Those researchers will pay between $1,000 and S$25,000 for the material. While the quality of the product is poor, it seems that the articles still manage to pass peer review.
Additionally, paper mills will pay publishers to accept their fake studies.
Between 2010 and 2020, the new tool revealed a 12 percentage point increase in the rate of potential fake papers published by some journals.
The nation with the highest number of potential fakes was China, contributing to just over half the red flags. Russia, Turkey, Egypt, and India were also significant contributors.
“Fake science publishing is possibly the biggest science scam of all times, wasting financial resources, slowing down medical progress, and possibly endangering lives,” researchers argue.
And the rise of generative AI such as ChatGPT only makes the scam more of a threat.
Critical research in all fields is being compromised thanks to AI-generated sham science paired with human greed. University research systems that reward publication quantity over quality are also contributing factors. Peers are also clearly not doing “peer review.”
I fear for the future of genuine science research that enhances our lives through knowledge, understanding, innovation, and progress.
We don’t need the Terminator from the future to kill us. It looks like AI is already doing that without time travel.
DONATE
Donations tax deductible
to the full extent allowed by law.
Comments
Maybe “publish or perish” is the way to root out the fake scientists (like Fauci/Fau xi).
The very foundation on which the publication of manuscripts is based is unvarnished, professional peer review.
When science became political, as it “climate change” and “gender affirmation/butchery,” and got published, there was no reason to assume that any paper was worth a grain of salt, hence the closing down of the big publication mills.
Unfortunately, bad science will thrive in a Wikipedia like review environment, and science will be gutted like everything else the left touches.
I would point out that there aren’t really any papers that contradict climate change and with respect to gender affirming care (particularly in children) it’s the lack of papers that’s the issue.
That said it is a big issue as a whole for science. Too many incentives to churn out any old shit. It’s sad that it’s now a profession investigating peer review integrity
“I would point out that there aren’t really any papers that contradict climate change”
You’re such a bald-faced liar.
Create an environment where someone submitting those types of papers would get punished by no grant funding and it would not be surprising to not see those papers. I would venture that some science disciplines have shot well past the level of damage that a Trofim Lysenko could even dream of. We are now deep into a cycle where public policy determines the research results, and not the other way around as originally intended.
Does this cast doubt on Doctor McGillicuddy’s MRNA COVID elixir?
Only at the Shady Eye Saloon with Hermione.
Lucky that elixir can still be used to remove stains from clothing, peel paint from steel, and cure gastrointestinal issues.
I have some in the back of my horse drawn cart, under the name of Dr. Percy Long’s Miracle Elixir.
It sounds a bit like Wiley’s Hindawi subsidiary is becoming the CNN of science journalism.
there is a proliferation of published science papers with the Climate science field being just one example. John Abraham is one who has published and or peer reviewed over 300 papers in just 15 years – that is close to 200 papers per year. Sounds very much like a paper mill
Another example is the recent study on 12% of asthma cases caused by gas stoves ( a meta study with zero actual field work). In my opinion, that study is borderline academic fraud, if not outright academic fraud.
Another more recent study had an estimate of 19,000 deaths per year in the US due to gas stoves. Again zero actual field work. A meta study using a review of 70 + prior studies, of which none of those original studies having attributed any deaths to gas stoves.
Also consider “baloney slicing;” where you take what would normally be a single paper, and break it up into parts that you submit as individual studies.
All to meet the “publish or perish” metric for success in science.
Climate scientists who delve into epidemiology are the worse agenda-driven idiots out there. The don’t confuse association/correlation with causality, they confabulate imagination with morbidity and mortality. It drives me crazy listening to some of the idiocy I read. When I was in the doctoral (not my MD, but my PhD) training, I had an excellent no-nonsense Hungarian biostatistician who was on my examination panel. She had no time for junk science, and loved to pull the legs off the climate alarmists’ insect body rhetoric. They spouted crap, but had no idea what it meant. Proved to me that their training in science was minimal, and most of it was spent on political spiel and not on sound science and supporting field/lab work.
This would be a lot more convincing if real world data hadn’t already confirmed the climate models are accurate
Funny how the proponents of the CLIMATE CRISIS!™ won’t openly and publicly debate anyone disagreeing with them. Instead, they ban, slander and censor any opposing viewpoints (Dr. Judith Curry comes to mind). That is behavior that suggests they know they will lose the argument. The CLIMATE CRISIS!™ is scientific fraud on a huge scale and is based on nothing more than money, power and control.
Bert E – the typical warming cycle and typical cooling cycle over the last 15k years has been around 200-300 years.
predicting the continuation of warming as the “climate models” have done during the middle of a warming cycle is not quite so impressive.
Sure. That’s why the models have to be frequently adjusted so they will “predict” the past few years’ data.
Climate “scientists” have been caught more than once altering data to fit their models and theories.
If the models were accurate, they’d have predictive power. None of their predictions that should have come to pass have come to pass. One could excuse that by saying that 20 or 30 years ago, when those predictions were made, the models weren’t as good as today. Even granting that, the best we can do is wait another 20, 30, or even 100 years before we can say that today’s models are “accurate,” because any predicted events must be so far in the future (because the climate doesn’t change overnight) that there’s no way now to gauge the accuracy of the models. Of course, it’s also likely that 20, 30 or 100 years from now, apologists for today’s researchers will say “The models weren’t so great back then. New climate models are definitely able to predict the climate 100 years out from now.”
Also, we’re not stupid. We realize that if 30 years ago a bazillion dollars was thrown at the alleged problem of climate change, when the predictions didn’t come true, said failure would have been attributed to the efforts to “fight climate change,” and would have been used to ask for even more money to “continue the fight.” There’s a reason why the whole thing seems like a scam.
BartE’s purpose in life is to waste our time.
You mean the altered data provided by NOAA, MetService, ACS, etc? All of which have been found to have been publishing false data. The ACS was most egregious, as it was unaware that the actual weather and temperature log as it was hand noted, was residing in the Australian Museum of Natural History in Sydney. When an intrepid researcher found the original volume, it was immediately noticed that The ACS had falsified the older “recorded” temperatures by making them considerably colder that were entered. NZ was found to have done the same to 2 of the 11 continuous reporting weather stations. When called on for an explanation, MetService invented the now infamous excuse that it was “norming” the data. As for NOAA, it is a joke. I have heard NOAA employees admit that the published NOAA temperatures are those of the hottest area instruments, usually at an airport. That merely going to a station in a relatively nearby field yields a temperature a full 2F lower.
Anyone with a home station or even access to a web site like Weather Underground will see that the actual temperature on the screen will be significantly lower than that published the following day.
it would be a lot more convincing if real world data hadn’t shown that all of their predictions turned out to be wrong. (fixed it for you)
Mathematical models provide only a forecasting tool, and they are subject to such great limitations that they are used in combinations with each other. See the hurricane models and how badly they predict the paths of the vast majority of hurricanes despite their years of “validation”. When you get into something as complex as biological systems, mathematical modeling becomes even more tenuous. Goito try another forced hockey stick?
“100% of the people who confuse causation with correlation will die.”
Omigod! I’m doomed!
So will 100% of those who don’t. They’re clever like that, they get you coming and going.
Error. Presuming the 15-year period is correct, the number of cases should be 3,000 or the per annum number of papers should be 20 to make the numbers work. Typos, most likely; it just looked funny. [3000/15 = 200 or 300/15 = 20]
yes should be 3000 / 15 years = 200 per year. that is not scientific research, but a paper mill
Surely you mean 20 papers per year…?
200 papers a year woul be 3000, not 300.
You may recall the time (2017-2018) when several scientists decided to send completely fake papers to demonstrate how the soft sciences are just a big scam (https://www.sciencealert.com/cultural-studies-sokal-squared-hoax-20-fake-papers). The best summary, written by the authors, is here (https://areomagazine.com/2018/10/02/academic-grievance-studies-and-the-corruption-of-scholarship/).
To wit: “The trio scaled-up operations to press out 20 hoax papers over a space of 10 months in parody of what they refer to as grievance studies. These were then sent to what they argued were the “best academic journals in the relevant fields”.
Most of the papers were inspired by some form of progressive political ideology, such as the role of patriarchy in modern society or the influence of imperialism.
“Sometimes we just thought a nutty or inhumane idea up and ran with it,” the writers explain.”
Sound familiar?
How about this from Slate, who tried hard to defend the grievance industry (https://slate.com/technology/2018/10/grievance-studies-hoax-not-academic-scandal.html):
“Going off these numbers, one might presume that “grievance studies,” as the sting would have it, is best defined as mostly feminism and gender stuff. That interpretation would seem to fit the authors’ prior work and interests too. This week’s project turns out to be the offspring of another hoax from Lindsay and Boghossian, published in May 2017. That one, called “The Conceptual Penis as a Social Construct,” was billed as an attack on gender studies in particular. “We suspected that gender studies is crippled academically by an overriding almost-religious belief that maleness is the root of all evil,” they wrote at the time. “On the evidence, our suspicion was justified.”
“Breastaurants?” “Dog on dog rape in dog parks?”
What about this: “Another tough one for us was, “I wonder if they’d publish a feminist rewrite of a chapter from Adolf Hitler’s Mein Kampf.” The answer to that question also turns out to be “yes,” given that the feminist social work journal Affilia has just accepted it.” (from the Aero citation, above)
Is there any reason to think these areas of “study” aren’t a complete scam on their own?
Even with improvements, AI produces data out of thin air and creates studies that don’t exist to back up its own assertions. I don’t buy the 90% accuracy comment, because that would have to be based on a study that may or may not be accurate in and of itself—e.g. a “systematic literature review” that often either excludes relevant studies due to investigator bias, or includes studies that are poorly done. As with all research, the problem is that humans are involved and we are flawed beings. The problem with AI involvement is that humans programmed AI and again—we are flawed—hence GIGO (garbage in/garbage out).
Lives are on the line in medical and clinical research. Look at the recent number of studies, particularly in oncology, that were retracted at Dana Farber Cancer Center due to repetition of manipulated imagery, and it will give you an idea of how deep the problem may go with academic reputations involved. Scary.
The 90 pct accuracy claim is just for a computer trained tool looking to check for AI use in a manuscript. It is not intended to be a bias or thoroughness check.
That said, based on what I’ve seen from most AI detectors, I still think that number is inaccurate.
Most likely it is inaccurate. A lot of that software was based, so I am told, on the old “academic integrity” software that my dissertation was run through several times at each point in the editing and commentary stage before final submission. AI was not in common use at all when I did my research, it was all old-fashioned hard work. Not a typewriter and seven bottles of Liquid Paper (if anyone even remembers that?!), but the equivalent. I don’t have much to say about AI because I’m leaving academia at the end of next month to get out of the DEI garbage. I’m only clinical faculty, but even we’re subjected to it. Contrary to what people read, AI doesn’t really make that good of a replacement for a well trained doc.
It doesn’t in the wrong hands….and handing it over to a new generation of medical grads that don’t know as much as their predecessors is akin to making someone an accountant because they are trained on a calculator but have no idea how to manually perform even basic arithmetic.
There has GOT to be a way for cagey septuagenarians to arbitrage their accumulated skills against incompetent/imposter Gen-Z/Millennials. We are missing an opportunity that my never occur again!
Only about 1 in 6 research grants that are deemed meritable are ever funded. Grade inflation has led to degree inflation where now even the most most mediocre are “earning” their Ph.D. and compare that to the 1980 when I was performing research in Chemistry where, in my experience, well less than half were capable of performing credible research. So take a glut of Ph.D.s that are mediocre to poor in quality, but still quite intelligent and high motivated problem solvers, put them in jobs where they either publish or they become unemployed, and see what happens. I left research in the early 1990’s due to the corruption and lies back then and its only gotten exponentially worse. Add to this the absence of ethics as witnessed by the COVID debacle, the idiocy of global warming, the gender silliness, and so much more, and I should think we’d be happy with the current rate of of fraud in science being so low.
Not to mention the intellectual incest prevalent in STEM.
Peer review is probably a net good but there are tons of biases involved and people who have built their career on a specific hypothesis will fight tooth and nail to suppress research that contradicts their claims, even if it is sound.
Once you get to a certain level, you can find flaws (or future directions) in anything if you look hard enough. Biased reviewers are known to be harder on contra research and suggest more studies needed to publish just to suppress stuff they don’t agree with.
healthy guy – a good example is from one of my clients. A researcher specializing in a few rare eye diseases, was dominated with research using mice. His study[ies] showed that rabbits were a better animal species to test than mice for those eye diseases. Major objections to his thesis. the unfortunate part was the success rate using mice wasnt very good, so there was little to lose by testing on rabbits.
There was a researcher at Tulane who as a peer reviewer sat on a paper of mine (which contradicted a viewpoint upon which their career was based) for over a year with a series of petty criticisms (series in that they were brought up one at a time at each submission). On a plus note, Katrina wiped their lab off the surface of the earth, so God works in mysterious ways, just saying. 😉
The best science right now is vulconology and astronomy, because Iceland and Webb are changing established “laws” every day. I have to humbly admit that my May 20 prediction for a serious resumption of the events on the penninsula was off. Having real-time GPS, web-cams, and earthquake data on-line has us peasants on the same playing field as the professionals, even from 5000 miles and 7 time zones away.
Science has always had a bit of political in it. My chronic pain work in the 1980’s was funded by NIDA even though there was little bearing to the drug abuse problems that they wanted to solve. But that is minor compared to the climate-change and harm-reduction efforts by “well-intentioned” but savagely ignorant liberals.
On a plus note, the voters in Portland demonstrated resistance to some of this malarky this week, with the Soros-backed DA soon to hit the pavement. Hopefully his opponent will make do on promises of getting serious with the miscreants. Of course, ALL of the bond measures went through, even if they only received 10% of the votes claimed. Money issues never fail here.
Also, there’s a bias insofar as what gets funded gets published (if only because unfunded work never gets off the ground). There is obvious bias in funding for certain fields (e.g. climate change) that necessarily tips the balance of what appears in journals. (And the problem compounds when the papers being submitted and published in those fields are junk.)
I recall a story some years ago of a grad student who wanted to study a certain species of squirrel in Colorado. He couldn’t get funding until he changed the direction of his study to “how climate change is affecting” the species he wanted to study. I’m sure readers here could recount similar stories.
In earlier generations, in other realms, the magic phrase was “Mao Zedong Thought.”
i recall a researcher trying to study a particular plant species (possibly agriculture plant) over the last 200-300 years, but get things to work out. Then discovered the primary reason was that the temperature record had been homogenized of that period. Made the study worthless. (my apologies since I dont remember the species – plant or even animal)
Yes, it is so political and keeps getting worse.
Back when I was naive enough to be on social media and expressing my opinions, I would openly explain all of the problems in current science to many people hoping that people would become aware and push for reform from outside (since it wasn’t going to come from within).
The wife of my PI as a post-doc (also a Ph D) messaged me and asked me what I was doing and tried to lecture me about how important it is that we tell the public about only the good that science does to increase support for funding research since funding lines are so slim. I blew her off….luckily, PI was professional enough to separate us.
By the way, these are problems that we all talked about openly and agreed on in our scientific circles and knew it was counterproductive….but mum is the word outside the bubble of science.
Hide the decline, baby!
AI…the name says it all. Synthetic awareness is a false statement.
You are somewhat right and somewhat wrong.
Synthetic consciousness is a theoretically possible concept that has not been achieved to date; AI is not self-aware.
You are right, though, that calling this stuff intelligent is a little marketing white lie. It should be called simulated intelligence or virtual intelligence.
Are we at the point where we can all acknowledge that peer review isn’t performing its role?
Papers must undergo serious review by actual experts for a rational person to assume their credibility. Those papers must be then replicated by another lab if we are to consider applying them in any context that matters.
Do those two things happen? Oh, they don’t? Then the scientific process cannot be deemed credible as it is currently practiced. Science doesn’t magically answer questions, but has to be done properly to arrive at reasonably probable conclusions.
Yep. As currently practiced peer review seems to have been abandoned in favor of a dynamic akin to the South Park Panderverse ‘put a chick in it and make it lame and gay’. So long as the paper’s title and summary promise to meet the demands of the woke orthodoxy then it slides right on through.
“Those papers must be then replicated by another lab if we are to consider applying them in any context that matters.”
Published papers are used by other researchers as starting point for data/result replication efforts. Not sure how one would go about reversing that process. The purpose of peer review is to weed out the obviously wrong, so that other researchers don’t waste their time/effort/money to demonstrate the obvious – that a published study was wrong. There’s not enough money to chase every paper with attempts at replicating the data of anywhere near every paper that’s submitted. Certainly few would be interested in doing so with regard to papers that haven’t yet passed peer review.
Well, yeah. Peer-review basically catches 1) Are appropriate methods used to determine the results? and 2) Does this stuff even make any kind of sense? I don’t expect peer review to catch measurement errors or liars, but they should weed out an awful lot of papers that end up published.
Replication of results is the only real check for honest measurement errors or just lies. If it isn’t replicated, then you have to take everything based on the skill/honor of the research team. That’s fine if it doesn’t matter. It’s not fine in anything that forms the basic for anything else or has real life applications.
Every time I see a statement like that, it brings to mind the decades in which the first investigator miscounted the number of human chromosomes, and decades worth of peer-reviewers and replicators dutifully reported the same (wrong) number.
“A man hears what he wants to hear, and disregards the rest…”
Like the republic if you can keep it….
Peer review is only as good as the integrity of the reviewers and the ability to put biases aside for objectivity. That erodes with each passing day.
This is exactly like anarchy. If a critical mass of the country suddenly decided to stop following any laws tomorrow, we would have anarchy in the blink of an eye. Law and order only works if the vast majority are willing to adhere to the code that benefits everyone else that also adheres to it.
the un controls america
only trump told them to gofthemselves and tried to do everything the right way but lefty /rino coalition stopped maga
fact!
Blaming AI is avoiding the real problem. I reviewed a paper the other day that may have been written with the assistance of AI but the much more important point is that the study it described was idiotic to begin with. It looked to me like it was designed by unusually stupid 7th graders.
When there are no standards, of course there will be fraud.
It would be interesting to know how many of these fraudulent papers comport with one of the globalist narratives such as Covid, climate change, gender nonsense, etc. The scientific process has been corrupted by The Science™ as promulgated by those who seek to rule us and control the purse strings of government grants. Ayn Rand was prescient in her portrayal of The State Science Institute as a corrupt actor in service of the looters and moochers in Atlas Shrugged.
I’m thinking along the same lines. See my comment above concerning what gets funded. If it’s the politically-correct “science” being funded, that’s what will end up published (legit or not). There’s bias in funding and bias in peer-review, and that’s a problem, because then it becomes difficult to falsify someone else’s research when that research has been done in a politically-protected (having guardians at both ends of the study & publish pipeline) field for an inability to get funded and/or published when the research is contrary to The Narrative™.
“Critical research in all fields is being compromised thanks to AI-generated sham science paired with human greed.”
If only there were some way to make better human beings.
John 3:16
If there was a way to make better human beings, god should have seen to it when he made them.
Babies are not born corrupt; they learn corrupt behaviors. Don’t blame God for that; if He forced you to be upright and ethical you would not have free will.
Nyet. Good behavior is learned. This is why children growing up without fathers are depicted in later chapters of Lord of the Flies.
Without God, we are animals.
Without fear of God, we behave like animals.
Without someone instilling the fear of God we have no reason to not be animals.
Job 11:12. Man is born like a wild donkey colt [and must be trained to be human].
“Babies are not born corrupt”
The doctrine of Original Sin was always a non-starter for me, but at least one mainline religion can’t do without it.
By the way, science is obviously not the only area affected. A lawyer was sanctioned by the bar not long ago for citing case law that didn’t actually exist. Where did he get it? From the hallucinations of an AI bot.
Yeah, that’s just laziness. Writing the brief himself was too much work; and so was vetting the AI’s results before submitting them.
Otter (AI version): “You fucked up. You trusted us.”
This is nothing new, fake “science” papers predate AI.
My uncle was a chemist at Dupont. They got the idea to buy the rights to a couple hundred papers, with the plan to use them as the basis for new products. So, they were cherry-picking the most likely candidates – and these were all in hard sciences.
Then they went to work, and quickly discovered they ought to have done their homework in advance: 99% of the papers were complete garbage. Not reproducible.
This was back in the 90’s.