Fake Science: Where is academic integrity heading?
Fake Science: Where is academic integrity heading?
Between the recent prison sentence of a prominent researcher, the $1 billion paper mill market and fake AI-generated data, academic integrity faces its biggest challenges yet. Here's how we fix it.
Last week, Dr. Paulo Macchiarini, a “pioneer” of stem cell research, was sentenced to over two years in prison for the undue experimental surgeries on three patients that led to unnecessary suffering before their deaths. However, the misconduct of Dr. Macchiarini and the resulting loss of life is just the start. The host institutions involved were reluctant to admit scientific misconduct and it took years before Dr. Macchiarini’s papers were retracted.
This incident brings light to the broader integrity issues pervading academic institutions and the scientific research community at large. The existing landscape of publish or perish mentality and need for recognition are compromising the very foundation of scientific research as a reliable and accountable system of expanding collective knowledge.
Here, we’ll cover the most prominent infractions of academic integrity today and what solutions are already underway to tackle them.
Academic integrity issues today
The pressure on researchers to publish is one of the key drivers in various disingenuous publishing schemes. Falsehoods surface in different ways, whether it’s gaming the publication system through paper mills, acquiring ill-earned citations, or simply making up research findings. Just as with any cat and mouse game, it is difficult to catch the culprits, and this one only gets harder with AI.
Fake citations, fake data, fake research
Sometimes fake research is obvious, but most of the time, it isn’t. Various factors can be duped, such as faking the data or bypassing real peer-review. Once published, additional fraudulent activity can occur, like getting more citations than deserved through citation hacking schemes.
Although the case of Dr. Macchiarini mentioned above sounds extreme, it’s not the first time. Another researcher, Dr. Hwang Woo-suk, an alleged “pioneer” in cloning, had a similar story of severe research misconduct. The main difference being that Dr. Hwang ended up cloning camels for United Arab Emirates royalty. Although these cases seem like outliers, they merely represent the more heavily publicised and penalised examples. A recent study estimated that 28%, or some 300,000 papers, in the biomedical field alone may be fake. Extrapolating this finding across science could mean half a million dubious papers per year (28% of the roughly 2 million papers published annually).
New technology, like generative artificial intelligence (AI), may only make the issue more challenging. Already, researchers are creating highly convincing, fake data, like fake AI-generated microscopy images. Not only are these images difficult to detect, but proving they are AI-generated is tricky too.
“The capacity of paper mills to generate increasingly plausible raw data is just going to be skyrocketing with AI,” - Jennifer Byrne, molecular biologist
Paper mills
Paper mills are arguably one of the most challenging issues within the existing research infrastructure. It’s a mechanism by which researchers get low-quality or dubious papers published by relying on centralised organisations, or other collaborating researchers. It’s not just morally and scientifically damaging, but expensive. Complicit researchers are estimated to spent 1 billion euros per year to have their authorship added to accepted papers. Not only are paper mills difficult to prevent (because they are motivated by the unavoidable pressure to publish) but notoriously hard to detect due to the disjoint nature of academic publishing via separate journals.
Researchers who participate in paper mills are difficult to distinguish. For example, Dr. Gunasekaran Manogaran was recently accused of running a potentially decade-long publishing scam which targeted the “special issues” of journals. Dr. Manogaran’s record, with several retractions involving fake peer reviews and rogue editors, doesn’t help his case. How long the scheme went unnoticed is a testimony to the difficulties in detecting and remedying paper mills. Only now people are catching on, after 10 years, 11,000+ citations, and enough papers to win Dr. Manogaran an h-index over 60.
Upholding academic integrity better
Significant literature is dedicated to the issue of plagiarism and integrity among students, but little focuses on the same issues among academics and professional researchers. Nonetheless, organisations, policies and initiatives do exist to try and remedy the pervasive issues plaguing the existing system. The majority of these solutions focus on publication process, trying to detect or altogether avoid questionable research from getting published.
Self-correcting solution: retraction
Academia does have self-correcting systems in place, although they often lack the support and transparency necessary to be maximally effective. One of the most important mechanisms is the retraction of papers. Unfortunately, various flaws and issues in execution mean it’s not effective as it could be. That’s partially why the Retraction Watch site exists, to shed light on existing retractions and boost transparency in the system.
One key issue with retractions today is that journals don’t react fast enough. In Dr. Macchiarini’s case, it wasn’t until this year that his 2015 paper (which was on the surgeries he was already guilty of misconduct for) was retracted, making the total count now nine. Journals need to react faster and clearly communicate retractions to ensure they no longer receive citations.
Fraud detectives
Dr. Adrian Barnett, a metascientist, recently brought to light a novel solution: research detectives, or individuals tasked to discover and expose faulty research. In fact, such detectives already exist such as the Dutch microbiologist, Dr. Elisabeth Bik, who has reported over 4,000 papers already. But such heroic endeavours are not intrinsically rewarded in the existing academic system. Dr. Bik does so from her own volition and seeks support via crowdfunding.
Thus, Dr. Barnett argues that specific funding may be allotted with the express purpose of cleaning out the fraudulent literature. An even more scalable approach, one that remedies the problem before it happens, involves randomly selecting paper submissions to undergo an additional check by the “fraud police”. Such an approach would not only work against fraudulent research, but also citation hacking and manipulated peer reviews. Although a novel approach, a big question mark still remains regarding who will fund and promote such work.
Organisations and standardisations
Several organisations have sprung up in order to tackle the issues of academic integrity pervading the publishing and research system. The Dutch nonprofit, STM Integrity Hub, offers a centralised hub and various resources including screening infrastructure and support for publishers. Their working groups provide best practices and guidelines, such as their report on AI Ethics in Scholarly Communication and Recommendations for handling image integrity issues. Their Working Group on Peer Review Taxonomy, since absorbed by NISO, created a standard for the peer review process to ensure a transparent and clearly communicated process.
The Committee on Publication Ethics (COPE) consists of individuals, publishers and universities worldwide, committed to creating an academic culture where ethical practices are the norm. They run a forum to discuss cases such as whether pseudonyms should be permitted or how to handle suspect fake images in publications. COPE’s key offerings are clear guidelines on how to handle potential cases of academic malpractice and courses on ethical publishing.
In addition to these organisations, several annual conferences focus on ethical research. Just next month, the UK will host the Research Ethics Conference (REC2023) and the European Conference on Ethics and Integrity in Academia (ECEIA).
The implications of these academic falsehoods, whether of prominent surgeons like Dr. Macchiarini or lesser-known researchers taking advantage of paper mills, extend far beyond academia. Not only do they compromise the scientific method, but further erode public trust in science. One of the key mechanisms in enhancing public trust rests in building trust with scientists themselves. Unfortunately, this is shot when academic infractions occur.
Can the scientific community come together to overcome the variety of pervasive threats to academic integrity? Although perfection is arguably not the goal, there is still plenty to remedy and many institutions and individuals are already stepping up to meet the challenge. In the end, however, integrity comes down to the individual.
What are your concerns regarding scholarly integrity today? Comment below or join the discussion by tweeting at us @LitmapsApp.
Resources
Research Ethics Conference 2023 - REC2023, 6 July 2023, University of Bath, UK
Retract or be damned: a dangerous moment for science and the public, June 22 2023
Scientific fraud is rising, and automated systems won’t stop it. We need research detectives, June 20 2023
AI intensifies fight against ‘paper mills’ that churn out fake research, May 31 2023
Fostering A Culture Of Trust: Insights On Academic Integrity And Research Ethics – Analysis, May 18 2023
Fake scientific papers are alarmingly common, May 9 2023
STM Advancing trusted research