The crisis in scientific publishing: from AI fraud to epistemic justice

There is a crisis in scientific publishing. Science is haunted. In early 2024, one major publisher retracted hundreds of scientific papers. Most were not the work of hurried researchers, but of ghosts—digital phantoms generated by artificial intelligence. Featuring nonsensical diagrams and fabricated data, they had sailed through the gates of peer review.

This spectre of AI-driven fraud is not only a new technological threat. It is also a symptom of a pre-existing disease. For years, organized networks have profited from inserting fake papers into the scholarly record. It seems that scientific publishing’s peer review process, intended to seek truth, cannot even tell the real from the fake.

These failures are not just academic embarrassments. In fields like global health, where knowledge means the difference between life and death, we can no longer afford to ignore them. Indeed, the crisis in scientific journals is not, at its heart, a crisis in publishing. It is a crisis of knowledge—of what we value, who we trust, and how we come to know. That makes it a crisis of education.

Crisis in scientific publishing: The knowledge we ignore

Consider what Toby Green has called the “dark side of the moon.” He is referring to the vast body of knowledge produced by established experts in international organizations. Volumes of high-quality reports and analyses come from organizations large and small. They contain immense expertise. Often, not only do they qualify as science. They may be more likely to shape policy and practice than most academic outputs. Yet this “grey literature” is rarely incorporated into the scholarly record. This is why Green is actively implementing projects to find, collect, and index such materials.

If the formal knowledge of some of the world’s leading experts is being left in the dark, what hope is there for the practical wisdom of a frontline nurse?

In the rigid hierarchy of evidence that governs global health, a randomized controlled trial sits at the pinnacle. At the very bottom, dismissed as mere “anecdotes,” lies the lived experience of practitioners. A nurse in a rural clinic who discovers a better way to dress a wound in a humid environment has generated life-saving knowledge that could be useful elsewhere. A community health worker who develops a sophisticated method for building trust with vaccine-hesitant parents has solved a problem in context. Yet, in our current culture, their insights are not data. Their experience is not evidence.

To dismiss such knowledge is an act of willful ignorance. Science, at its best, is a process of disciplined curiosity. Its fundamental purpose is to reduce ignorance and expand our understanding of the world. To willfully ignore entire categories of human experience and expertise is therefore a betrayal of the scientific ethos itself. It is an active choice to remain in the dark.

Crisis in scientific publishing: the architecture of exclusion

This devaluation of practical knowledge is not an accident. It is a feature of a system designed to exclude. The modern ideal of science began with a radically open mission. As the scholar John Willinsky has meticulously documented in his history of Western European science, the creation of scientific journals in the 17th century was intended to create a public commons of knowledge, accelerating progress for the benefit of humanity. The principle was one of access. How was this mission corrupted?

The architecture of modern science was built on a colonial foundation. Its violence was not only physical but also scientific and intellectual. Frantz Fanon, the Martinican psychiatrist who became a theorist of decolonization in the crucible of Algeria’s war of independence, described colonization’s deepest work as the effort to “empty the mind of the colonized.” This is a systematic process of convincing people that our own histories, cultures, and ways of knowing are worthless.

Generations later, the Māori scholar Linda Tuhiwai Smith detailed how this was put into practice. She showed that Western research methodologies themselves were often not neutral tools of discovery but instruments of empire. The acts of observing, classifying, extracting, and analyzing were used to control populations and invalidate their knowledge systems, replacing them with a single, supposedly universal, European model of truth.

This worldview pretends to be a neutral, “view from nowhere,” a concept also critiqued powerfully by the white American feminist philosopher Donna Haraway. She argued that all knowledge is situated—shaped by the position and perspective of the knower. You see the landscape differently from the mountain top than you do from the valley. A complete map requires both perspectives.

Echoing this, her philosophical and geographical sister Sandra Harding argued that by excluding the perspectives of marginalized people, dominant science becomes “weakly objective.” It is blind to its own biases and assumptions.

Crisis in scientific publishing: Fear of knowledge

A common and deeply felt fear among scientists is that embracing diverse forms of knowledge will lead to a dangerous relativism, where objective truth dissolves and “anything goes.”

Harding’s work shows this fear to be misplaced. She argues that the “view from nowhere” provides not a stronger, but a more brittle and fragile grasp of the truth. A truly “strong objectivity,” she contended, is achieved by intentionally seeking out multiple, situated perspectives. This does not mean that all views are equally valid. It means that by examining a problem from many standpoints, we can triangulate a more robust and reliable understanding of reality. We can identify the biases and blind spots inherent in any single view, including our own.

This process is the antidote to the willful ignorance mentioned earlier. It strengthens our grasp of objective truth by making it more complete and more honest.

Can change be paved by good intentions?

Today, the need for a change in research culture is widely acknowledged. The world’s largest research funders publish reports calling for more diversity and inclusion. Yet we observe paralysis rather than progress. The individuals who sit on the decision-making committees of such institutions will almost certainly not fund a project with a primary investigator whose work is not validated by the existing system of prestigious but exclusive journals. Elite global scholars leading the vital movement to “decolonize global health” first established their legitimacy by adhering to conventional norms, then began using the master’s tools to have their critiques of the system heard. Such contradictions illustrate how deeply the exclusionary norms are embedded.

Since top-down change is caught in such contradictions, a meaningful path forward may be to change the culture of science from the ground up. The core challenge is to correct for epistemic injustice: the wrong done to someone in their capacity as a knower. This injustice takes several insidious forms.

The most obvious is testimonial injustice. Imagine the scene. A senior male doctor from a famous university presents a finding and is met with nods of assent. His words carry the weight of evidence. A young female nurse from a rural clinic presents the exact same finding based on her direct experience. Her knowledge is dismissed as a “story” or an “anecdote.” She is not heard because of who she is. Her credibility is unjustly discounted.

Even deeper is hermeneutical injustice. This is the wrong of not even having the shared language to make your experience understood by the dominant culture. The community health worker who builds trust with hesitant parents may have a brilliant system, but if they cannot articulate it in the formal jargon of “implementation science,” their knowledge remains invisible. They are wronged not because they are disbelieved. They are wronged because the system lacks the concepts to even recognize their wisdom as knowledge in the first place.

Projects like Toby Green’s grey literature repository or initiatives like Rogue Scholar, pioneered by Martin Fenner, that assign a permanent Digital Object Identifier (DOI) to science that was not previously in the scholarly record, are practical interventions. But this not a technological problem. It is an educational one. Changing a culture that perpetuates these injustices is the primary work. Within this larger project, new tools can serve as tactics of resistance. As such, they can be used to support acts of epistemic defiance, for example by creating a formal, citable record of knowledge that exists outside the traditional gates. Yet they remain tools, not the solution.

The science of knowing

You cannot fix a broken culture by patching its systems. You must change its DNA. The crisis haunting science is not ultimately about publishing, fraud, or peer review. It is a crisis of education—not of schooling, but of how we come to know. If physics is the science of matter, education is the science of all sciences. It provides the architecture of assumptions and values that shapes how every other field discovers and validates truth.

A new philosophy of education is needed, one that includes these three principles:

1. It must recognize that the most durable knowledge comes from praxis—the cycle of acting in the world and reflecting on the consequences.
2. It must be built on collaborative intelligence, understanding that the most difficult problems can only be solved by weaving together many perspectives.
3. It must pursue strong objectivity, not by erasing human perspective, but by intentionally seeking it out to create a more complete and honest picture of reality.

To change science, we must change how scientists are taught to see the world. We must educate for humility, for critical self-awareness, and for the ability to listen. This is the work of creating a science that is not haunted by its failures but is directly contributes to a more just and truthful account of our world.

References

1. Boghossian, P., 2007. Fear of knowledge: Against relativism and constructivism. Clarendon Press.
2. Couch, L., 2021. Wellcome Diversity, equity and inclusion strategy [WWW Document]. Wellcome. URL https://wellcome.org/what-we-do/diversity-and-inclusion/strategy (accessed 11.8.22).
3. Fanon, F. (1963). The wretched of the earth. Grove Press.
4. Fenner, M., 2023. The Rogue Scholar: An Archive for Scholarly blogs. Upstream. https://doi.org/10.54900/bj4g7p2-2f0fn9b
5. Gitau, E., Khisa, A., Vicente-Crespo, M., Sengor, D., Otoigo, L., Ndong, C., Simiyu, A., 2023. African Research Culture – Opinion Research. African Population and Health Research Center, Nairobi, Kenya. https://aphrc.org/project/african-research-culture-opinion-research/
6. Green, T., 2022. Wait! What? There’s stuff missing from the scholarly record? Med Writ 31, 44–48. https://doi.org/10.56012/ajel9043
7. Haraway, D. (1988). Situated knowledges: The science question in feminism and the privilege of partial perspective. Feminist Studies, 14(3), 575–599. https://doi.org/10.2307/3178066
8. Harding, S. (1991). Whose science? Whose knowledge? Thinking from women’s lives. Cornell University Press.
9. Smith, L. T. (2012). Decolonizing methodologies: Research and indigenous peoples (2nd ed.). Zed Books.
10. The Social Investment Consultancy, The Better Org, Cole, N., Cole, L., 2022. Evaluation of Wellcome Anti-Racism Programme Final Evaluation Report – Public. Wellcome, London. https://cms.wellcome.org/sites/default/files/2022-08/Evaluation-of-Wellcome-Anti-Racism-Programme-Final-Evaluation-Report-2022.pdf
11. Wellcome Trust, 2020. What researchers think about the culture they work in. Wellcome, London. https://wellcome.org/reports/what-researchers-think-about-research-culture
12. Willinsky, J., 2006. The access principle: The case for open access to research and scholarship. MIT press Cambridge, MA.

Image: The Geneva Learning Foundation Collection © 2025

#AIFraud #DEI #diversityAndInclusion #DonnaHaraway #epistemicInjustice #epistemology #hermeneuticalInjustice #LindaTuhiwaiSmith #philosophyOfScience #RogueScholar #SandraHarding #scientificPublishing #strongObjectivity #testimonialInjustice #TobyGreen

Il nuovo disordine mondiale / 8: mai più per un pugno di conchiglie https://www.carmillaonline.com/2022/03/22/il-nuovo-disordine-mondiale-8-mai-piu-per-un-pugno-di-conchiglie/ #fardellodell'uomobianco #Controinformazione #dedollarizzazione #economiamonetario #Emiratiarabiuniti #AlessandroOrsini #globalizzazione #scambioineguale #BrettonWoods #NazioniUnite #Interventi #Afganistan #conchiglie #ElSalvador #Occidente #SudAfrica #TobyGreen #FakeNews #Mariupol #Pakistan #petrolio #Zelensky #bitcoin

By John Helmer

We’re in a world where people don’t really understand what they want until you put it in front of them,’ says Toby Green Head of Publishing at OECD. He’s talking about the challenge of creating new digital products in a technology landscape that is changing very quickly (with no end to the ‘technology treadmill’ in sight) and where market research is of limited value; where what happened in the past in educational publishing is a poor guide to what will happen in the future.

This reflection comes from looking at OECD’s markets, which span both higher education and the workplace, and a remit that embraces not only information dissemination but, to a degree, instruction. We’re talking convergence.

Toby Green will chair the plenary session on ‘Cross-fertilisation’ at the ALPSP International Conference. The convergence of the education and workplace learning markets is likely to be a theme for this session, so we took the opportunity to convene a three-way discussion involving Reda Sadki, a learning innovation strategist who is working with OECD on precisely this area.

We discussed drivers for convergence, some of its effects, and also opportunities and threats for publishers.

Moving beyond a dissemination mindset

Reda’s vantage point on this phenomenon of convergence is informed by his time at the International Federation of Red Cross and Red Crescent Societies (the IFRC), where he pivoted from managing publishing to ‘learning systems’. The IFRC, he says, was an organization that published massive amounts of information (750 information products, 12 million printed pages in 2009), with “little measurable impact”. ‘Ultimately I came to the realisation that the value in what was being published by the world’s largest humanitarian network could be found in the instructional and training materials, with a global audience of 17 million Red Cross and Red Crescent volunteers. Where you could find impact was in the publications that teach people in a humanitarian emergency how to do very basic things such as putting up a tent and providing first aid care.’

He characterises the transition this realisation prompted as being from a concern over maximising dissemination – counting eyeballs and downloads – to looking at a deeper kind of impact in terms of what was happening behind the eyeballs. It is a shift that he implies publishers need to make themselves if they are to capitalise on the opportunities offered by this convergence.

Drivers of convergence

Reda sees two fundamental shifts driving convergence.

One is about changes in the economy of effort to do certain things. Publishing starts with dissemination and under the traditional model would tend to stop at that. It doesn’t necessary look at look at what people are doing with what it disseminates – largely because, pre-internet, it would have been uneconomic to do so. Technology has lowered the cost of, for instance, collecting rich data about what people are doing with a particular piece of knowledge.

The other is about the changing nature of knowledge itself. The book gave us a ‘container’ view of knowledge, where now – with knowledge flows getting faster all the time – it looks more like a process than a product. Attempts to capture and compartmentalise knowledge are doomed to fail, in his view, as they do not provide the answers that we need to be able to provide it in any useful way. Being an expert today is much more about knowing where and knowing how than it is about the individual accumulating large amounts of knowledge.

Echoing Reda’s first point, but framing it in a perhaps broader context, Toby sees the appearance of new possibilities for action with the advent of digital as the decisive factor. ‘If you think of the offline world, on both the publishing side and the education/training side, there were some natural constraints to what you could do …’

The book (or textbook, or journal) was bound. It had a finite number of pages and could be shipped to only so many people. The classroom could only have a finite number of people in it, and was very difficult to scale without massive expense in both infrastructure and people (i.e. teachers). Online removes a lot of those scaling constraints; so a class that could previously only reach 30 people can now reach hundreds of thousands.

Online has also massively lowered the cost of updating published information. A new print edition of a textbook, for example, is a major undertaking. In the offline world updates to knowledge would happen in batches, because it wasn’t feasible to do it in any other way. Online allows you to have a rolling update – giving us the concept of a living book – or, equally, a course that is constantly being tweaked and kept up to date.

These changes allow new ways of thinking. There are significant changes to the old paradigms – but they are changes that a lot of people are still trying to get used to, both on the education side and on the publishing side.

One area that publishing has been very successful in, Toby feels is integrating technology with content, and he gave several examples of workflow tools such as Mendeley that bear this out, and the work of other players in the wider information industry such as Bloomberg and Reuters.

However going beyond these essentially resource-based models and becoming more instrumental in the process of learning is another matter, and considering this led us to look at the different cultures these converging (or colliding) industries have.

Culture and authority

One of the most beautiful things about publishing, in Reda’s view, is the way in which culture, in both the specific and the wider senses of that word, is embedded in its fabric. This gives a different feel for the value of the content, and its importance in terms of the emotional relationship we have with works of the mind and aspects such as cultural diversity in what is published. While e-learning taps into a rich history of learning theories and education, it still has something to learn, he feels, from the culture of publishing in this respect.

Knowledge management, by contrast – which he feels to have failed – seems obsessed with putting pieces of data into pigeonholes, without proper regards to the more important activity of building a culture to make sense of the vast amounts of information and data that organisations receive and generate.

From the publishing side, Toby observed that the linkage of education and training has always been weak. Textbook sales were seen as by-product of publishing activity, where existing titles were picked up on by educators – or else the preserve of a highly specialised branch of publishing that knew how to do them.

Now, with the collapse of barriers that limited thinking in the offline world, and with digital reducing costs and lowering barriers to entry, the idea of publishers working with partners to adapt their content to create courses is far more achievable. And here is a further cultural change: the idea of working with partners. ‘Before, companies did everything themselves; they didn’t really use networks of freelancers and partners in the way we do now’.

My own reflection on the different cultures, having worked in e-learning and digital publishing, is that there is less concern about provenance of knowledge on the training side of the fence. Academic publishing has a culture of sources, citation and reference that is currently in the process of automating in a characteristically rigorous way (CrossRef, ORCID, etc.). In e-learning, on the other hand, where content is often produced using an organisation’s internal SME knowledge, individual authorship tends to be more submerged, and it is often possible to wonder: where is this point of view coming from; who is telling me this?

As somebody who works for a ‘who’ (the OECD) Toby can’t help but believe that at the point of convergence, this difference offers an opportunity for organisations like his own whose content carries the stamp of accepted and established authority in their particular field. This could also apply to the learned societies, but doesn’t necessarily hold true for larger, more generalist commercial publishers.

Effects of convergence, chilling and otherwise

Given the way that internet power laws operate in any online space – tending to favour one or a very few brands and condemn everyone else to place on the ‘long tail’, these questions of identity and authority are critical online. Certainly their effects have been seen in the case of MOOCs.

Arguably, it is the presence of educational ‘super-brands’ such as Harvard and Stanford that has allowed online education to break through to public consciousness in the way it now has, under the banner of MOOCs. Interestingly however, other HE institutions in this rarified upper strata that have chosen not to participate in this gold-rush so far – notably Oxford and Cambridge in the UK – don’t seem to be especially troubled by the phenomenon.

It is the ‘squeezed middle’ of second tier universities who see MOOCs as a threat to their livelihood, and the opinion of many is that solution in future will be for institutions to find or build specialisms in particular unique areas. Get ‘niche’.

Reda locates a particular opportunity here in the troubled issue of ‘the fit in today’s world of the capacity of universities to prepare people for the workforce or for the demands of society’. Sub-degree, competency-based qualifications represent, in his view, ‘a huge gaping hole’ that knowledge-producing institutions are in a privileged position to address.

He cites a client he worked with who had seen an Oxford University course on the area they worked in, but believed they could themselves build one ‘a hundred times better’. This sparked for him the idea that an organisation that has the practice – that actually does the job – could now, through the affordances of technology, build an educational offering of high quality.

An organisation that in addition starts with a strong publishing function is particularly well placed since they will already have the quality development processes that will make it much easier to build educational experiences around that content.

Playing the long game

Of course, underlying all this talk of opportunities is the necessity for publishers to make their digital investments pay, and while moving into creating educational experiences around content might represent an opportunity for some organisations, there usually has to be some threat element in play to compel action.

Reda pointed to the scrabble for data around MOOCs, which as early as 2013 prompted publishers to offer access to their textbooks within MOOCs in return for the user data. In a data-driven world, he would consider not having some such access to this type of data as a risk.

This has to be see in the context of attempts by publishers to use digital to bring textbooks to life, not all of which have proved wildly successful with users, and the idea, argued by some, that MOOCs themselves are textbooks: that, ‘MOOCs perhaps represent the first form of digital textbook to reach a mass audience’.

Given factors like these, organisations can’t afford to not experiment and try new things if their businesses are to grow and survive.

In Toby’s view, publishers still largely think they’re in the business of selling content. He sees very few examples of textbook publishers migrating online in a way that works. ‘Part of the challenge is that since individuals are so reluctant to spend any money for content online – and bearing in mind that the offline textbook market was largely an individual-purchase model – it is very hard to see how a textbook publisher is going to get a return if they simply put their textbook online’.

Data driven-models mean that money is made elsewhere than in the same transaction, so the challenge is to look at your publishing business in the round. A publisher such as Wiley, whose acquisitions in the learning space follow a strategy around the lifetime value of a customer – from education through to their professional life – might (notionally) balance losses in one part of the business by larger gains in another. This would involve looking at the value of the individual rather than the value of the training.

‘That’s what makes the web so hard, but at the same time so interesting: you have to consider where the value is, and the lifetime value could be very long … it’s very difficult to look individually at each particular piece: you have to look at it holistically.’

https://redasadki.me/2014/08/12/convergence-and-cross-fertilisation-semantico-talks-to-toby-green-and-reda-sadki-about-publishers-and-learning/

#digitalTransformation #JohnHelmer #learning #OECD #publishing #Semantico #TobyGreen