Journalology #132: Trust processes not people
Hello fellow journalologists,
I attended the STM meeting on innovation and research integrity this week. One of the key learning points for me was that publishers are thinking about ‘trust’ differently compared with a decade ago. Traditionally, academic publishing has been a trust-based system: editors take authors’ claims at face value; editors are not the police.
However, that stance has had to change in recent years because of the proliferation of fraudulent content from paper mills and low quality AI-generated slop. Now, publishers need to trust their systems and processes, rather than their authors, in order to identify bad actors at source. I wrote about this on Linkedin on Thursday and there was a healthy debate in the comments section that you may want to read or contribute to.
News headlines
NIH’s proposed caps on open-access publishing fees roil scientific community
Sometime next year, the National Institutes of Health (NIH) will announce new limits on how much funding grantees can spend on publication fees to make their articles open access, or free to read. The agency says it aims to reduce “unreasonably high” article-processing charges (APCs), a goal many researchers embrace. But public comments released last week from more than 900 researchers, academic institutions, and publishers reveal deep concerns about a proposal that one commenter, radiologist Geoffrey Young at Mass General Brigham, calls “well-intentioned, but misguided.”
JB: You may remember that back in July the NIH issued a Request for Information on Maximizing Research Funds by Limiting Allowable Publishing Costs. The deadline for responses was September 15 and this week the collated responses were made available — all 1389 pages of them. Thankfully, Christopher Steven Marcum has created a machine readable version of the comments here.
The Science news story notes that:
Other options are still under discussion at NIH and among publishers. One would provide authors and their institutions more flexibility by capping reimbursements for APCs at 0.8% of a research grant’s direct costs—the agency’s calculated average of what researchers currently use to pay these fees—or $20,000 over the length of the award, whichever is greater. Some publishers generally opposed to the idea of NIH-enforced APC caps have reluctantly endorsed this approach as the least objectionable option.
This approach may encourage authors to publish fewer, better, papers, which would be a good thing. Journals that allow authors to write very long papers (I’m thinking of Cell in particular) may become more appealing (to authors if not to readers).
APC pricing discussions often centre on commercial publishers’ profit margins. However, society publishers are concerned about APC price caps too. Here’s an excerpt from the comments left by a group of influential clinical societies (American Society of Clinical Oncology, The Endocrine Society, American Thoracic Society, American Society of Anesthesiologists, American Heart Association, American Academy of Neurology, American Society of Nephrology, American Society of Hematology, and perhaps some others — the list looks as though it was truncated).
We firmly believe that our society members should be free to publish the results of their work in the journals that are most appropriate. For some papers, it might be a general medicine journal; for others, it might be a basic research or subspecialty journal. Either way, there will naturally be wide variations in publishing models. Some journals, such as leading subscription-based titles, do not charge authors at all, ensuring that publishing decisions are not contingent on an author’s ability to pay. Other titles offer open access options, where fees vary. The NIH is limiting the academic freedom of researchers to choose the most appropriate venue for sharing their research results.
It’s worth noting that three of the leading US society journals — Science, The New England Journal of Medicine, and JAMA — do not offer a paid open access option. That’s probably because the APC price would need to be very high in order to cover their costs; it’s not in their commercial interest to offer hybrid OA. An APC price cap may encourage other journals to either remove their APC option entirely or to lower costs — by being less selective or by reducing editorial quality control — in order to reduce the APC price to meet the NIH threshold, whatever that turns out to be.
Springer Nature retracts, removes nearly 40 publications
The papers attempted to train neural networks to distinguish between autistic and non-autistic children in a dataset containing photos of children’s faces. Retired engineer Gerald Piosenka created the dataset in 2019 by downloading photos of children from “websites devoted to the subject of autism,” according to a description of the dataset’s methods, and uploaded it to Kaggle, a site owned by Google that hosts public datasets for machine-learning practitioners. The dataset contains more than 2,900 photos of children’s faces, half of which are labeled as autistic and the other half as not autistic.
JB: My jaw hit the floor when I first read about this dataset. Here’s the backstory:
The dataset first came to Springer Nature’s attention last month through separate investigations into two papers, Kersjes says. The research integrity team was about to start investigating one “article of concern” when Guillaume Cabanac, professor of computer science at the University of Toulouse, alerted the team to the other one, which contained tortured phrases—strange phrases used in place of established ones, a possible sign that the text was generated by artificial intelligence.
So, if I understand this right, the Springer Nature research integrity team was alerted to the problem on November 16 and within less than a month had retracted 5 articles with 33 more retractions planned. The papers should never have been published, of course, but let’s give the research integrity team some credit for acting so quickly. As Tim Kersjes, Head of Research Integrity at Springer Nature, wrote on LinkedIn:
But this was one of those cases where working in research integrity makes a difference. Cases like this made me realize why working in research integrity can be so rewarding: retracting these articles is important. The dataset was unethical and obviously flawed and shouldn’t have permeated the literature as it had, so cleaning it up is the right thing to do.
Springer Nature was not the only publisher affected. The Transmitter, which broke the story, has identified 90 articles that use this unethical dataset. You can see which journals and publishers are affected here: Elsevier, Frontiers, IEEE, MDPI, JMIR, PLOS, Taylor & Francis, and Wiley all appear on the list (plus many other smaller publishers). Springer Nature has shown that rapid action is possible. Will the other publishers be willing and able to follow suit?
China leads research in 90% of crucial technologies — a dramatic shift this century
The ASPI’s Critical Technology Tracker evaluated research on 74 current and emerging technologies this year, up from the 64 technologies it analyzed last year. China is ranked number one for research on 66 of the technologies, including nuclear energy, synthetic biology, small satellites, while the United States topped the remaining 8, including quantum computing and geoengineering. The results reflect a drastic reversal. At the beginning of this century, the United States led more than 90% of the assessed technologies, whereas China led less than 5% of them, according to the 2024 edition of the tracker.
JB: Yes, I know I sound like a stuck record, but China is the future. Ignore that fact at your peril.
A new preprint server welcomes papers written and reviewed by AI
At most scientific publications, papers co-authored by artificial intelligence (AI) are not welcome. At a new open platform called aiXiv, they are embraced. The platform goes all in on AI: It accepts both AI- and human-authored work, uses built-in AI reviewers for baseline quality screening, and guides authors through revisions based on the chatbots’ feedback.
JB: Who will foot the electricity bill for this vAInity project? An AI-designed perpetual motion machine?
UK university sector reaches deals with four major publishers
The UK higher education sector has agreed terms on deals with four out of five major publishers. Jisc, the sector’s technology and digital organisation, has been locked in negotiations with Elsevier, Springer Nature, Taylor & Francis, Wiley and Sage since March 2025 to secure new agreements “that reflect the priorities identified by the sector”, including advancing open access. On 12 December, Jisc said that “sector-agreed thresholds for acceptance” had been met by Taylor & Francis, Elsevier, Springer Nature and Wiley, following consultations with institutions. Agreement with Sage is to be considered later because the publisher did not meet an October deadline for consultation on its offer with the sector. Jisc told Research Professional News that an agreement with Sage is currently under consultation and “we expect to inform institutions of the decision on this offer by 19 December”.
JB: This all sounds very positive, but some UK institutions will likely not renew:
With the UK’s higher education sector facing major financial pressure, there is widespread expectation that not all institutions will be able to afford to renew their deals, and Jisc said that acceptability of the agreements “ultimately lies with institutions”.
Over the past month I’ve been having my own negotiations with medium and large scholarly publishers. Before we move on to the news headlines, I would like to express my gratitude for the support of the five publishers (including 2 of the Big 5) that have purchased access to Journalology for everyone in their company; for the many other smaller organisations that have set up group subscriptions for a portion of their workforce; and for the fellow journalologists who have subscribed individually.
Just shy of 15% of the recipients of this email will be able to read all the news below the fold as a result of a paid subscription. Your support means I’ll be able to spend more of my time on Journalology in 2026. I’ve got lots of exciting things planned for next year. Please consider upgrading your subscription.
Keep reading with a 7-day free trial
Subscribe to Journalology to keep reading this post and get 7 days of free access to the full post archives.