Scientific publishing needs urgent reform to retain trust in research process

The Guardian

20-07-2025

The dysfunctions of scientific publishing that your article so aptly captured derive from two forces (Quality of scientific papers questioned as academics 'overwhelmed' by the millions published, 13 July) – researchers are incentivised to publish as much as possible and publishers make more money if they publish more papers.
Artificial intelligence will not fix this. Churning out more papers faster has got us to this place. Given current incentives, AI will mean churning them out even faster. A paper written by AI, peer-reviewed by AI and read only by AI creates a self-reinforcing loop that holds no real value, erodes trust in science and voids scientific inquiry of meaning. Research is driven by our wonder at the world. That needs to be central to any reform of scientific publishing.
Instead, the driving forces can be addressed by two measures. Incentives for researchers can and should prioritise quality over quantity, and meaning over metrics. And publishers' extortionate fees (fuelling profits of more than 30%) can and should be refused by those who pay them. Both the incentives and publishers' contracts are governed by the funders of research – universities, research councils and foundations. Their welcome attempts to engage with these problems through Plan S, which aims to make research publications open access, have not succeeded because these have been captured by publishers that twisted them to their advantage, making yet more profits.
There are examples, often beyond the global north, of scientific publishing that is not geared towards generating profits for publishers. SciELO (which is centred on Latin America) is one, and the Global Diamond Open Access Alliance champions many others. We have much to learn from them. Research is in a parlous state in the English-speaking world – at risk for the truths it tells in the US, and for its expense in Britain. Funders have the power radically to alter the incentives scientists face and to lower the rents extracted by publishers.Dan BrockingtonIcrea (Catalan Institution for Research and Advanced Studies)Paolo CrosettoGrenoble Applied Economics LaboratoryPablo Gomez BarreiroScience services and laboratories, Kew Gardens
Your article on the overwhelming volume of scientific papers rightly highlights a system under pressure. But the deeper dysfunction lies not only in quantity, but in the economics of scholarly publishing, where publishers cash in on researchers' dependence on journals for academic careers. The academic publishing market systematically diverts public research funds into shareholder profits.
Open access was meant to democratise knowledge, but its original vision has been co-opted by commercial publishers. It was BioMed Central (now Springer-Nature) that first introduced the 'author pays' model to secure revenue streams. With article processing charges (APCs) now being the dominant open-access model, authors routinely pay between £2,000 and £10,000 to publish a single article, even if the cost of producing it does not exceed £1,000.
Some of us attended the recent Royal Society conference on the future of scientific publishing, where its vice-president, Sir Mark Walport, reminded the audience that academic publishing isn't free and that if we want to remove paywalls for both authors and readers, someone must pay the bills.
We argue that there is already enough money in the system, which allows leading publishers such as Elsevier to generate profit margins of 38%. Our most recent estimates show that researchers paid close to $9bn in APCs to six publishers in 2019-23, with annual amounts nearly tripling in these five years. These most recent estimates far exceed the $1bn estimated for 2015-18 that your article cites.
As further emphasised at the Royal Society meeting, publishers monetise the current role that journal prestige plays in hiring, promotion and funding. Therefore, in order to make open access sustainable and to put a stop to these extractive business practices, it is essential to reform academic assessment and decouple it from knowledge dissemination.Stefanie HausteinAssociate Professor, School of Information Studies, University of Ottawa; Co-director, Scholarly Communications LabEric ScharesEngineering and collection analysis librarian, University Library, Iowa State UniversityLeigh-Ann ButlerScholarly communication librarian, University of OttawaJuan Pablo Alperin Associate professor, School of Publishing, Simon Fraser University; Scientific director, Public Knowledge Project
Academic publishing is creaking at the seams. Too many articles are published and too many journals don't add real value. Researchers are incentivised to publish quantity over quality, and some journal publishers benefit from this. This detracts from the excellent, world-changing and increasingly open-access research that we all need to flourish – and that quality publishers cultivate.
Generative AI only scales up these pressures, as your article shows. Something has to change. That's why Cambridge University Press has spent the last few months collaborating with researchers, librarians, publishers, funders and learned societies across the globe on a radical and pragmatic review of the open research publishing ecosystem, which we will publish in the autumn.
Focusing on generative AI or on low-quality journals alone is insufficient. We need a system-wide approach that reviews and rethinks the link between publishing, reward and recognition; equity in research dissemination; research integrity; and one that takes technological change seriously.
The system is about to break. We need creative thinking and commitment from all players to fix it and to build something better.Mandy HillManaging director, Cambridge University Press