Csaba Szabo is not convinced with all those announcements, manifestos, white papers and declarations on how to fix scholarly publishing. Therefore, he offers a more radical solution.

Here some past declarations:

• The Budapest Declaration, for Open Access (1994 and 2002)
• The Berlin Declaration, for Open Access (2003)
• The Singapore Statement, on research integrity (2010)
• The San Francisco Declaration, on Research Assessment (DORA, 2012)
• The Leiden Manifesto, for research metrics (2014)
• The Sarajevo Declaration, on integrity and visibility (2016)
• The Hong Kong Principles, for research integrity (2020)
• The Barcelona Declaration, for open research information (2024)
• The Paris Declaration, on forensic scientometrics (2024)
• The Stockholm Declaration, for reformation of science publishing (2025)

Many more declarations on scholarly publishing here. And because Csaba is professor at the University of Fribourg in Switzerland, I decided to call this “The Fribourg Declaration”. Needless to say, the following views are Csaba’s own and most likely not of his employer.

---

Beyond Papermills: Why Scientific Publishing Must Be Rebuilt from the Ground Up

By Csaba Szabo

The papermill crisis is not an external attack on science but a mirror held up to its dysfunctional ecosystem. The for-profit publishing industry, posing as a newly found saviour of research integrity, is in fact its chief beneficiary and enabler. If science is to recover its credibility, the system that feeds on its dysfunction must be dismantled. Not repaired, not polished: dismantled.

The existence of papermills — shady operations that fabricate scientific papers for paying “authors” and guarantee their publication — has been known for years (STM & COPE report 2022, Bishop & Abalkina 2023. Abalkina et al 2025). What’s new is the attention they are getting in the mainstream scientific literature. People who hear about them for the first time are amazed when they realize the sheer volume of papermills’ activity. The brazen arrogance of their operations. Their infiltration into every layer of academia. The staggering sums of money flowing through their networks. While it is not difficult to infiltrate their circles to expose their modus operandi (as I did, see here and here), the criminals behind the papermills remain untouched, hiding in plain sight — protected, intentionally or not, by the very system that pretends to fight them. 

The Publishers’ “War” on Papermills: Virtue-Signaling and Hypocrisy

The publishing establishment “declared war” on papermills. United2Act, as they announce with banners and hashtags and zoom presentations. Their newest invention, the “Trust Marker,” promises to flag “authentic” work, based on certain characteristics of the authors and institutions — unconnected to whether the paper’s claims are true or reproducible. The Lancet has just assembled a new committee, in conjunction with the so-called “World Conferences on Research Integrity (WCRI) Foundation”; they will meet at their upcoming meeting in Vancouver in May 2026. Even the controversial publisher MDPI announced to start using “AI-powered image proofing” to identify fake images. Most recently, the “Stockholm Declaration” by Bernhard Sabel and Dan Larhammar alerted everybody that “paper mills’ increasingly pollute the scientific literature with fake articles reporting fake data” and called on everyone to “implement legislations, regulations and policies to increase publishing quality and integrity”.

These campaigns look righteous at first glance. But they are hollow, hypocritical, and self-serving. It is marketing, not reform. WRCI conferences have been going on since 2007, without any visible impact. There is an endless supply of integrity and metascience conferences, workshops and a similarly never-ending list of studies, guidelines, checklists and concordances –– some of them academia-initiated, many of them supported by the publishing industry. But let’s be clear: this industry is not the victim in the story of papermills — it is an accomplice.

Based on the massive volume of fake articles published every day, papermills represent a significant revenue stream for the publishing industry. The “top” journals can afford to be selective, but even the publishers of the most reputable journals have established lower-tier journals, to beef up their income. In other words, every publisher which relies on a revenue, benefits from the flood of garbage science.

The gray-zone and predatory journals have always happily taken papermill submissions to collect the publication fees, and the whitelisted Open Access journals are not much better. And, of course, closed-access journals need to fill their pages with something in order to collect the subscription money from academic libraries. Unless something drastic happens, the publishers will continue with their current practices –– while applauding their own “anti-papermill” initiatives.

At least in part based on information from For Better Science, a recent analysis published in PNAS (Richardson et al 2025) shows that papermills operate hand in hand with corrupt editors — even in supposedly reputable journals. A few names were recently exposed in Nature, again based in part on information from For Better Science. Some editors quietly resigned –– or were removed. For every corrupt editor caught, dozens more remain. The system tolerates them because large parts of it depend on them.

As I have discussed in my book, while the amount of grant funding for research has stagnated over the last decade, the number of published papers has increased dramatically. I am afraid, the most plausible explanation is this: a lot of these publications are not based on actual research.

 

Papermills Are a Symptom, Not the Disease

Every time a new detection tool appears, papermills adapt. They moved from copy-paste fabrications to AI-generated text and images, and now to algorithmic exploitation of open-access databases. Each crackdown simply drives evolution. 

There will be always plenty of publishers ready to take the papermill money. This feels like a never-ending process. And pursuing the papermillers themselves? Bound to fail – they operate offshore and in the dark web. The US FTC “won” its case against the OMICS Group years ago — and OMICS is still alive and profitable. 

Papermills exist because the academic system demands them. They are not the disease — they are a symptom.

The real problem is the demand. The desperate, constant demand from people — let’s stop pretending they’re scientists — inside universities who need papers to survive, advance, or impress. Why do they buy fake science? Because their institutions reward publication volume and citation metrics above everything else. Until these perverse incentives are dismantled, papermills will thrive. When the demand dies, the supply will die also. Everything else is virtue signalling.  

Beyond Papermills: The Deeper Rot

Papermills are just a visible fungus growing on a rotten tree. Systematic replication studies have already shown that 80% of biomedical papers (or more!) — even those in “top” journals, even when repeated with assistance from the original authors — fail to replicate (Szabo 2025a, Szabo 2025b). Most of this irreproducibility cannot be blamed on papermills.

Irreproducible papers come from mainstream labs, with grants, prestige, and institutional blessing. The empirical truth is that the biomedical scientific literature, as a whole, is profoundly unreliable.  Papermills are merely the black-market version of the same sickness: a culture that rewards output, not validity; novelty, not truth; publication, not knowledge. Fixing this requires a total systemic reform — of training, funding, data analysis, and quality control. Until that happens, we are not curing the disease.  

Why the For-Profit Publishing System Must Go

The for-profit publishing industry does a great job in presenting itself as the only possible mechanism for the communication and distribution of scientific results. However, historically, scholarly communication and the publication of scientific and academic work did not rely on a shareholder-owned, for-profit business model (Stephen Buranyi in Guardian, 2017). University presses, learned societies and academic communities took responsibility for dissemination and quality control. For much of the twentieth century journals were owned or managed by academic societies (often non-profit) and libraries subscribed to them; the service model was part of the academic infrastructure rather than a commercial venture. For-profit intermediaries gradually gained dominance, but there is no intrinsic reason why the publication of scholarship must be structured as a profit-maximising business rather than as a public good.

The latest “papermill crisis” and the self-serving and inefficient “declaration of war” on them is just the latest proof for the obvious mismatch between the logic of profit and the logic of scholarship. The system today imposes high costs, exhibits growth in volume rather than demonstrable improvement in quality, fails to guarantee the reliability of the published claims, and shows no evidence of genuine self-reform. The economics of current academic publishing is a staggering waste of public resources.

Despite the fact that public and charities pay for most research, peer-reviewing is carried out largely by unpaid academics (Aczel et al 2021), and the final product is then sold back to institutions (and indirectly governments) via subscriptions or article processing charges (APCs). Publishers are consistently making profit margins of 30–40%. The academic sector (i.e., ultimately, the taxpayers) ends up paying dearly, and largely for a system that redistributes publicly-funded knowledge through a commercial intermediary whose margins are high and whose value added is questionable. 

The “brand” value of journals is increasingly inflated; the journal impact factor and prestige hierarchy drive researcher behaviour (to publish in high-impact titles) rather than the intrinsic value of the published work. Which, by the way, keeps decreasing, in terms of each published paper’s real-word impact. But when the system incentivises volume and prestige rather than replication, rigour and quality, what is achieved is rather an expansion of output, not an expansion of trusted validated knowledge.

This is the only reason why papermills exist: to fill the growing gap between the amount of available real research output and the virtual pages that an ever-growing number of for-profit journals must fill up with something. Almost anything. 

Gradual Improvements Are Hopeless

There is a lot of talk about making various improvements of the current system. Improving the peer review system (Waltman et al 2023, Margalida & Colomer 2016. Pinfield et al 2014, Aczel et al 2025). Awarding trust markers (COS 2025, Butler 2025) and other distinguishing features to “legitimate” groups of scientists. Improving the policing of fake images and AI-generated text. The list goes on. As long as the for-profit industry can pay lip service to such improvements, they will be embraced.

An often proposed and rather popular alternative suggests that scientists should “reclaim the ownership of journals” and place them in the hands of academic societies or universities. This suggestion, while made in good faith, would never succeed, because it would leave the underlying incentives, such as the metrics-heavy academic evaluation system, and the overall hypercompetitive publish-or-perish mentality, untouched. “Society journals” are in the money-making business, too: in this case the money is being generated for the society which owns them.

If a system –– any system –– does not work, if it does not serve its original purpose, it must be drastically reformed, or, better yet, it should eliminated and replaced. Stopgap measures will not work: the only solution is the urgent and complete abolition of scientific journals in their current form.

The Way Forward: We Must Put Scientific Publishing in the Hands of Funders 

In a new system, the current journal-based model of scientific result dissemination would be completely abolished (or, better said, it would cease to exist, on its own, because it would no longer be fuelled by the endless stream of manuscript submissions). Instead, funders would take direct responsibility for scholarly communication and data dissemination. They would establish an open infrastructure — preprint servers, community-managed peer review platforms, data-code repositories — and decouple certification (peer review, replication) from branding and business. In this model:

1. When an investigator (or group of investigators) is funded, they would have to agree that the funders’ platform will be the sole means of data dissemination and publication.
2. The funders would provide quality assurance, quality control and data integrity support: in other words, they would become active stakeholders, partners in the actual research.
3. Once a significant project component is completed, the grantees would post their findings (all raw data and their analysis) on a funder-mandated open archive. The data would be accompanied by an abstract and a manuscript, which is then immediately made public and collated by abstracting systems (e.g. PubMed)
4. Post-publication commenting (we don’t even need to call it peer review) would only begin at this point; both by funder-appointed panels and by direct input from the scientific community; comments would be open, signed or transparent.
5. Independent replication studies by third-party investigators must be encouraged, and performed in a blinded and controlled (industry-standard) manner. Independent replications would have to be part of the condition of each grant: separate funds would have to be reserved for this sole purpose. The results of the replication efforts would be posted on the funders’ server –– linked to the original body of data and manuscript. Studies that pass independent replication would receive a “stamp” of certified reliability.
6. Only those investigators who produce valid and reliable data would receive further funding. Institutions whose investigators generate large amounts of irreproducible or fraudulent data will be banned from further grant support, forcing them to improve their quality control processes.

This approach aligns publishing with the public-funded nature of research, removes the profit layer, and ties the system to community governance. This system, by the way, is not new. A decade ago the Wellcome Foundation has attempted to come up with something similar (“Wellcome Open Research”), but it failed, because –– after a public uproar –– it did not mandate its grantees to publish in its own journal.

The reform suggested above will only work if it is mandatory. The benefits would be massive: reduced costs, increased access, stronger quality control (via replication), realignment of incentives toward reliability rather than novelty.

Conclusion: Papermills Are Not the Problem — The System Is

The for-profit scholarly-publishing system with its self-serving and vacuous declarations to “fight” papermills has outlived its usefulness and lacks a credible mechanism for self-reform, because its structure is founded on profit rather than scholarship. It must go. Gradual improvements at the edges of the system do not work, otherwise they would have worked by now. 

Dismantling the current system and shifting publishing control to funders and research institutions is the only route to a more sustainable, scholarly-centric, transparent and cost-effective scientific ecosystem. Needless to say, any reform along the suggested lines will be difficult, because the publication industry, which is deeply entangled with the current academic system, controls much of the public discourse and is very good at self-preservation. Also, the current granting agencies are typically led by academic scientists who are also embedded in the current, dysfunctional scientific and publication ecosystem. So they must go as well. The management of the granting bodies should be replaced with a new crop of government administrators and reform-minded scientists whose sole interests are quality, reproducibility, validation, and the eventual translation of the findings for public benefit. 

It will be very hard to force a true reform. But it must be done. There is no other way forward.  

---