In April 2025, Csaba Szabo participated as session chair at Fostering Accountability for the Integrity of Research Studies (FAIRS), a research integrity meeting in Oxford, which was organised by Dorothy Bishop. This is his conference report, as a guest post.

The FAIRS presentations were published in PDF form on OSF, unfortunately there are no video recordings. I will provide my own opinions on what to foster at the end. Here I would like to briefly mention that I was not invited to speak at the FAIRS meeting. Moreover, although I was previously asked to fill out two long questionnaires as prerequisite for panel participation, I soon understood that any attempt of further contribution from my side would have, uhm, created tensions. Therefore, I chose not to invest money into participating as a guest, not even online.

---

FAIRS: “Frustrating Attempts at Integrity Reform in Science”

By Csaba Szabo

Between the 7th and 9th of April, Prof. Dorothy Bishop and Prof. Jaideep Pandit (both Fellows of St John’s College, Oxford, UK) organized the FAIRS meeting at their own alma mater. FAIRS, of course, stands for Fostering Accountability for the Integrity of Research Studies –– as opposed to my invention in the title of this report. It attracted about 80 in-person participants and about 120 online attendees. The audience consisted of academic scientists interested in research integrity, journal editors, publisher representatives, university research integrity officers, sleuths, some students, and a few people from the media. The entire program is available on the meeting website.

Dorothy Bishop, of course, is not only a highly accomplished (now-retired) scientist specializing in developmental neuropsychology but also a long-time science integrity warrior. She is probably the most academically accomplished scientist who devotes her time to this extremely important but entirely thankless and immensely frustrating issue. She used to be a Fellow of the Royal Society (until she resigned in protest against the election of Elon Musk into this society – a fascinating story, but one that does not belong in the current article). What does belong, however, is that Dorothy has been doing all the right things for scientific integrity reform for at least the last decade, including sleuthing, blogging, vlogging, tweeting, raising the issue of scientific misconduct and the replication crisis in dozens of interviews, testifying before the UK Parliament, and now, finally, organizing a 3-day meeting.

In the spirit of the day, I loaded all presentation PDFs from the meeting into ChatGPT and asked it to extract the most commonly used words to produce a Word Cloud, which can be admired here. (Each word’s font size is proportional to the frequency of the word used in the presentations.) All of these words are, of course, entirely familiar to the readers of For Better Science. Some words that are sorely missing, however, are “reform,” “policy,” “penalties,” “consequences”, and “solutions.” More on that later. First, I will try to summarize the main topics discussed at the meeting.

Preclinical Scientific Misconduct  

A significant portion of the talks focused on various aspects of scientific misconduct and research integrity, illustrating diverse issues including fraudulent data, problematic peer review, authorship conflicts, plagiarism, and the ethical dimensions of research practices. The topic of data fraud and manipulation was covered by several speakers, including the world’s most prominent expert (and recent Einstein Foundation awardee), Elisabeth Bik. She highlighted the extensive issues of manipulated images and falsified data that For Better Science readers know all too well, emphasizing the problem of extremely slow responses from journals and institutions.

A particularly sobering realization for me occurred when she showed AI-generated histological images that were indistinguishable from real ones. She stated that Western blots can now also be created this way. If this is indeed the case, the usual methods of fraud detection (visual observation, ImageTwin, etc.) will soon become obsolete, and nobody currently knows how AI-generated (i.e., unique) images will be detectable in the future. (At the meeting, there were many discussions and even breakout sessions on how to address the threat of AI; however, I did not hear any solutions beyond broad statements like “we ought to fight AI with AI.”)

Ivan Oransky provided further insights into the mechanisms of retractions and discussed their consequences for researchers, emphasizing that sanctions vary widely, including (usually slight) career penalties and (usually rarely enforced) legal repercussions. He also highlighted how journal practices, citation impact, and university rankings incentivize misconduct, calling for a reassessment of current publishing and ranking systems.

Nick Brown illustrated how statistical tests such as GRIM (Granularity-Related Inconsistency of Means) can detect impossible numerical values indicative of fraudulent research. He and other statistical experts who attended the meeting were adamant that truly random and “life-like” numbers are difficult to fake. They stressed that detailed statistical analysis can detect much fraud and should therefore be performed by journals at the very first step of the evaluation process, i.e., prior to sending the paper to reviewers. Such analysis can, of course, be most effective if all raw data are included as part of the submission.

The topic of institutional response and accountability was also covered by multiple speakers. For example, Raphaël Lévy presented his own horror story, conservatively titled “Inadequate Protection of Whistleblowers in France,” outlining the various retaliatory consequences he suffered for several years after exposing a set of fraudulent papers published by someone at his own institution.

Several speakers, including Anna Abalkina and Guillaume Cabanac, discussed the rapidly growing problem of paper mills. Cabanac provided illuminating and humorous examples of the “tortured phrase” problem, which occurs when fraudsters try to bypass plagiarism detectors by mindlessly substituting synonyms.

My own take on paper mills and tortured phrases is that — while there is no question that every fake or plagiarized paper contaminates the literature — most scientists can immediately identify these paper-milled “products,” and many of them already know how to handle them (directly into the trash bin). While reading tortured phrases can also be entertaining, as soon as scientists encounter a phrase like “pinnacle flag to clamor proportion” and recognize it as a distorted version of “peak signal to noise ratio,” they — after they stop hysterically laughing — know exactly how to deal with that paper (once again, straight into the trash bin). So yes, let’s clean up the literature, but I think it is immensely more important to catch the “big fish,” i.e., high-stature investigators who serially publish problematic data. Such cases can mislead entire fields of research and lead to the waste of many millions in research funding, both in direct grant money spent on hopeless directions and in opportunity costs. I’m thinking here of fields like stem cells in myocardial infarction (Piero Anversa), beta-amyloid research in Alzheimer’s disease (Eliezer Masliah), and others. To me, this is where the real benefit lies — in preventing such “runaway trains” from leaving the station. Exposing these issues is much harder and far more challenging, both methodologically and, when the affected bigwig decides to sue the sleuth, legally as well.

Clinical Research Misconduct  

Obviously, when misconduct occurs in a clinical area, it can directly affect patient health and safety. Although the entire sad Paolo Macchiarini saga is well-known to all For Better Science readers, it was still a chilling experience to hear Patricia Murray discuss her personal involvement with the case. What I find worth stressing is that Macchiarini, of course, did not work in isolation, yet nearly all of the individuals in Sweden and the UK who played various roles in enabling, supporting, and/or protecting Macchiarini remain active and unpenalized.

In another talk, the legendary British cardiologist and whistleblower Peter Wilmshurst recounted 40 years of personal experiences investigating research misconduct, including high-profile cases involving medical fraud. Once again, his presentation highlighted the persistent lack of adequate institutional responses.

When Nancy Olivieri gave her presentation, the same themes of institutional self-protection, pharma company corruption, and retaliation against whistleblowers were re-emphasized.

My immediate takeaway after the clinical session could be summarized by the following exasperated thought: “What are we all doing, getting hung up on flipped Western blots and duplicated histological images in obscure research papers, when some people are getting away with actual murder?!” (More on this later.)

Cultural, Social, and Geopolitical Dimensions 

Several speakers discussed how cultural and geopolitical forces intersect with scientific integrity. Till Bruckner‘s talk on the “weaponization of integrity measures” illustrated how scientific integrity measures can be misused politically, economically, and culturally, raising the risk of collateral damage to innocent researchers. Along the same lines, Stephan Lewandowsky discussed the challenges posed by fraudulent research to the credibility of scientific literature, emphasizing the balance between rigorous accountability and overly aggressive policing. (My two cents: we are so far from overly aggressive policing that I really do not see this as a serious concern. I wish we were already at the point where we had to worry about this.)

Patricia Kingori analyzed the phenomenon of “shadow scholars,” primarily from Kenya, who are hired to produce fraudulent academic content, primarily for American and British college students. This talk made me wonder: if a significant proportion of college students become accustomed to cheating on their essays even before beginning, say, a PhD in a biomedical field, and if their experience teaches them that cheating is an effective way to get ahead without being caught—what does that teach them for the rest of their careers?

Li Tang presented an update on China’s massive investment in research and highlighted some of the country’s challenges regarding scientific integrity reform. It was clear from her talk that funding agencies are well aware of China’s severe scientific integrity problems and that systemic changes are necessary to curb research fraud. However, it was less clear to me what the necessary next steps should or will be. For what it’s worth, she insisted that the famous “cash rewards for publications” system no longer exists in China, and stressed that granting agencies now focus on the quality rather than the quantity of publications, which should have some mitigating effect on paper mill product “consumption.”

The Official Consensus

As expected, there was extensive discussion around enhancing transparency, strengthening institutional accountability, protecting whistleblowers, reforming publication practices, adopting technological innovations, and recognizing the broader social and political contexts that shape scientific misconduct. However, in addressing these challenges, the conversation did not move beyond what I call the “standard solutions” in my book Unreliable — that is, ideas and initiatives which are palatable to official stakeholders but have neither worked nor will ever really work. Nevertheless, I list them below, using the solemn, official wording typically employed, which makes them appear more presentable but not one iota more implementable:

• Strengthening governance structures in academic and funding institutions.
• Developing advanced statistical and digital tools for detection and prevention.
• Advocating stronger protection mechanisms for whistleblowers.
• Promoting widespread adoption of open science practices.
• Reforming academic incentives and publishing standards to reduce misconduct opportunities.

Some Take-Aways

To be honest, I’m not sure what I expected from the meeting. I was hoping to get an update on the state of the “Research Integrity Field,” which I certainly did. I was also hoping to meet some of the main figures in the field, which I did as well—both those who are well-accepted by the establishment, like Dorothy Bishop, Ivan Oransky, and Elisabeth Bik, and many sleuths whose “official status” is less recognized. I also had the pleasure not only of meeting significant whistleblowers such as Patricia Murray, Peter Wilmshurst, and Nancy Olivieri, but even chairing their session. All of these individuals came across as serious, dedicated, and deeply concerned.

At the same time, it simply does not feel right that exposing bad science, forcing retractions, and triggering (rather sporadic and highly variable) consequences rest in the hands of private individuals — many of them working under pseudonyms and without any kind of legal protection. The mere fact that this is how the system currently operates is itself a problem. Moreover, I did not hear any plans or assurances from any stakeholder at the meeting indicating that this sorry state of affairs will change anytime soon. Academic institutions, publishers, and grant-awarding organizations remain largely uninterested in retrospectively cleaning up the literature, leaving this task to the sleuths — a real “David vs. Goliath” scenario (and, quite often, a Sholto David vs. Goliath scenario…).

At both the plenary sessions and small-group discussions, I kept asking attendees about their ideas for possible reforms and potential solutions. To my amazement, most were unprepared to discuss this matter. (It’s also possible that everyone at the meeting — except me — had long ago concluded that the system is FUBAR: there are no solutions, and they didn’t even understand why I was trying to raise the issue. But for a moment, let’s assume solutions might exist.) From the responses of those participants who did offer opinions, a few things became clear.

First, the clinical and preclinical sides of the problem should be separated. On the clinical side, there are already plenty of rules, regulations, IRBs, disciplinary boards, and legal frameworks; thus, the solutions must largely focus on better enforcement of existing processes. On the preclinical side, however, it appears that even the legal and regulatory framework is murky, down to fundamental questions like, “Is the production of fake data an illegal activity?” or “Is the attempt to publish false data a criminal offense?” or “What would happen if someone walked into a police station and reported a scientific misconduct-related matter?” Many attendees supported the idea of setting up a separate regulatory body or even a specialized police enforcement unit (by analogy with existing units for art or financial fraud). And (almost) everyone agreed that serial or severe misconduct should, at the very least, result in revocation of the individual’s academic credentials — similar to how, after repeated violations, people can lose their driver’s license.

Second, there seems to be significant disagreement among various stakeholders about the focus of reforms. “Establishment” players (representatives of funding agencies and academic integrity offices) prefer to emphasize prevention: better training, improved data collection, and better archiving practices. In contrast, the sleuth and whistleblower communities clearly advocate stronger penalties and even the introduction of new (legal) categories to clamp down on severe and repetitive scientific misconduct — even if it “only” involves preclinical/basic research.

As expected, I remained unimpressed by most members of the publishing industry I met at the meeting. Representatives of major publishing houses tried to assure me that they’re doing everything possible to improve research integrity and reject substandard or fraudulent papers. They claimed to be “totally open and eager to introduce new integrity measures,”…. provided these measures don’t affect their profit margins. I spoke with many academic journal editors who seem to have good intentions, yet many of the journals they manage still do not employ full-time statisticians, do not require the submission of raw data, and openly acknowledge that even if they catch a problematic or fraudulent paper, it will eventually end up published somewhere else.

But the group that frustrated me the most were the representatives of the “establishment.” Prior to attending this meeting, while writing the final chapter of Unreliable, I spent considerable time thinking about the various stakeholders and their roles. I concluded that, inevitably, some percentage of individual investigators will cut corners or engage in questionable research practices (or worse), given that the pressure is immense, the incentives substantial, the chance of getting caught low, and the penalties minimal even if caught. I also concluded that academic institutions have numerous conflicting interests, such as securing grants and protecting their reputations and these interest will severely limit their interest in true integrity reforms. (Indeed, many universities seem to place useful idiots into their research integrity offices.) Moreover, it is plain to see that the publishing industry, being profit-driven, remains more interested in profits than fundamental reforms. Thus, by elimination, I concluded that if reform is ever to occur, it must be driven from the funding side — imposed on the entire “system” by grant-awarding organizations and their governments.

After the three days of the FAIRS meeting, however, I am less hopeful. Representatives from the official side offered weak guidelines, non-binding “concordats,” and — in some cases — blatant cop-outs. More “coordination is necessary”, more “meetings are necessary”, “we must act gradually”, “we are doing all we can”, “you can already see our progress on our website”, etc. One representative of a British funding agency, during a small-group discussion, engaged in striking “whataboutism,” stating, “Look around, there are all sorts of problems, cheating, and misconduct in all walks of life; for example, look what’s happening in the Catholic Church.” I never thought I’d see the day when priests’ kiddie-fiddling would be cited as an official justification for a lack of comprehensive scientific reforms. In sum, I went into the meeting believing that grant-awarding agencies might be part of the solution; after all, if anyone, they should care where their money goes. Now, I fear that instead of them being part of the solution, they might instead be part of the problem, just like the other “stakeholders”. But, to be fair, there were relatively few official representatives present at the meeting — an issue also pointed out by Ivan Oransky, who advocates incremental rather than radical approaches but still feels that government and funding agency representatives must be convinced to initiate some reform.

Little Rays of Hope

There were a few bright spots presented here and there. Nick Brown described various methods (e.g., GRIM) to detect numerically impossible research findings, and he hinted at the existence of many other such techniques. Similarly, John Carlisle explained several rigorous statistical methods for detecting fraud in clinical trial datasets. If such methods were broadly implemented by journals at the screening stage (at least by those journals that actually care about what gets published), some progress could be made; many problematic papers could be spotted and eliminated right at the outset of the refereeing process.

But — as mentioned earlier — if one journal rejects a problematic paper, another one, typically a lower-tier journal but quite possibly still an indexed one, could (and probably will) publish it instead. One journal editor explained to me that the publishing industry had begun creating a networked (black)list of problematic papers and bad actors, along with a warning system through which all journals could be alerted when a fraudulent submission appears. However, it seems a class-action lawsuit halted this initiative. One can only hope that this legal matter will soon be resolved, as a journal-wide warning system could significantly reduce the tsunami of bad scientific papers that we should brace ourselves for as the AI revolution advances.

Uri Simonsohn (aka one-third of “Data Colada”) explored statistical methods to detect fraudulent results and introduced a practical system (“AsCollected”) aimed at preventing misconduct by ensuring data integrity and (hopefully) reproducibility via data tracking and preregistration. He did not have enough time to explain the entire system in detail but mentioned it will be piloted by a social science journal, with the current setup mostly tailored to social science-type studies. While I can see how the system might work with formalized protocols such as clinical trials or social-science studies, I suspect it would require significant tweaking to make it similarly useful for fast-moving, rapidly changing preclinical research, where experimental designs frequently evolve based on week-by-week data collection.

Finally, Liz Perkins (who holds the strange title of “Named Person for Research Integrity” at the University of Liverpool) presented her university’s efforts to combat research misconduct. Her talk went well beyond platitudes like better governance, accountability, and education and described concrete processes that are already in place, as her university is taking the matter extremely seriously. (Admittedly, the system in Liverpool was established in response to their infamous Dan Antoine/HMGB-1 diagnostic assay fiasco — but better late than never, I suppose.) I particularly liked their routine use of “random spot checks” of laboratory data and research practices — as opposed (or in addition) to the usual reactive processes triggered by whistleblower reports or PubPeer comments.

There was one additional bright idea mentioned during a discussion, made by Nick Brown. He suggested that funding agencies should establish and operate their own journals, mandating that their grantees publish there. Such a system, of course, would not charge APC fees, could be connected to open databases and even to pre-registration of the studies and could perhaps eliminate some of the temptation and hubris associated with publishing in “luxury journals.”

In Conclusion…

So, after spending three days at the FAIRS meeting, what have I learned? First of all, if each time somebody mentioned the “tip of an iceberg”, I would have collected 5 quid, I would have left Oxford with a wad of cash. Instead, I left Oxford with the realization that scientific integrity control feels like a never-ending game of whack-a-mole — except the moles are actually ‘crafty’ scientists, criminal publishing gangs, massive institutional blind spots and a publishing system powered by profit margins and wishful thinking. Sure, we have heroic whistleblowers and brilliant sleuths working in their free time, but real institutional accountability remains depressingly absent. I heard about some neat tools like statistical fraud detection and a few useful ideas like data audits of randomly selected research groups, but let’s face it: these are band-aids on bullet wounds while the big fish continue swimming untroubled through the system. Unless governments and funding agencies stop hiding behind toothless guidelines, non-binding concordats seeking to “support” research integrity (support? how about establish and then enforce!), and spectacularly creative excuses (“Hey, at least we’re not as bad as the Catholic Church!”), meaningful reform seems about as likely as publishers volunteering to cut their profits. Until then, the integrity crisis continues, with whistleblowers getting sued and fired and fraudsters getting promoted. The rest of us are left to laugh bitterly at tortured phrases and ridiculous excuses, while we are bracing ourselves for the upcoming tsunami of AI-generated scientific garbage.

In the meantime, the “science integrity meetings” will continue. For me, this was the first one I ever attended, and I found it extremely useful: it brought together representatives from multiple sides of the issue, and the discussions were honest and stimulating. I was quite surprised to discover that there has been an entire ongoing series of meetings nominally dedicated to scientific integrity since 2007—apparently without much effect. There will also be two more “integrity” meetings this year alone, one in Croatia and another in Australia, and the “official” series will continue next year at yet another fancy location, this time in Vancouver, Canada.

I only wish future meetings would place more emphasis on policy changes and action plans rather than becoming mired in definitions, guidelines, and endless committee deliberations. I also believe the presence of more governmental representatives would be an important addition to future conferences, so they can see first-hand the state we’re in, understand what’s coming, and realize how quickly the situation could deteriorate from bad to disastrous if significant reforms are not implemented. Perhaps this would serve as a wake-up call, finally steering them in the right direction.

---

Coda by Leonid Schneider

Since this is my own website, I can shoehorn here my own opinions without anyone putting their foot down and showing me out.

Now, I do agree with Csaba about the pointlessness of these stakeholder discussions on research integrity. Except that their pointlessness is the actual point, similar to industry’s greenwashing of their climate and environment damaging products: everyone can continue with wasteful and destructive activities while having a good conscience.

I do disagree that the way to go is to convince scholarly publishers and academic stakeholders about the need to crack down on science fraud. They are grown-up and financially very successful people, so it is very unlikely they waited for you to educate them with your PowerPoint presentation about the right way of thinking.

Publishing executives only care about revenues and profits, and those are tied to reputation, and this is exactly where publishers can be hit and made to cooperate. In fact, they only became United2Act on Asian papermills because of the reputational damage from the scandals exposed by my sleuthing colleagues. And even there, publishers try to titrate the right amount of retractions vs inaction, because of course their industry would immediately collapse without the papermills.

Academic stakeholders, i.e. executives of universities, research institutions and funding agencies, also know which side their bread is buttered. Even if they are not research fraudsters themselves – they arrived into their top jobs only because of professional networks. Which quite often include science cheaters, academic bullies and harassers, and their abettors. Let me remind you of just two members of the Senate of the German Research Council, DFG:

If you wants something to be done against smoking, you don’t preach to Philip Morris, no matter how clever and science-proof your arguments are. You will only succeed by appealing to the public and its elected representatives, directly and via the mass media. It is the same for science fraud.

Sleuths I work with do the most important task: they find fraud, and record it on that immensely revolutionary website, PubPeer. You can help make their findings reach the public and politics. If you are a university professor, your status elevates you and your opinion will be heard, all you have to do is to try contacting the media or elected MPs, and to alert them to the true state of science in your home country.

Most (if not all) high profile research fraud scandals only exploded with the consequences to perpetrators because of the coverage in mass media. Once it’s in the news, universities and journals have no other choice but to act on fraud. If as a journalist you want to wait out for academic “due process” to run its course (to the predictable whitewash), there will be very little or more likely nothing at all for you to write about.

Even though my website reaches all the right people, it alone cannot generate the required impact from the society and politics, for that the PubPeer evidence needs to be picked up by mass media. And the journalists will only report about science issues when professors advise them. Again, ask Simone Fulda.

Too often instead, journalists, especially science journalists, see themselves as cheerleaders for science and will never write a bad word about a scientist, certainly not about an elite scientist, because this is what the antivaxxers and fascists do, right? Actually, even antivaxxers and fascists listen to scientists, it’s just that they have their own “scientists” to follow.

Others report about retractions, For Better Science causes them. By the way, why don’t we have an automatic retraction database still? The Retraction Watch Database is still compiled entirely manually, despite having been sold to Crossref. Maybe YOU, dear reader, can code something?

Every one of you has much more talent, power and influence than you think you have. If you need a role model, look up to Patricia Murray or Elisabeth Bik.

Research integrity is a grass-root effort, a bottom-up approach where all of us can contribute, one way or another. No change, however incremental, will happen because some senior exec saw the light after attending a research integrity lecture.

It’s up to YOU, do something.

---