If your week flew by — we know ours did — catch up here with what you might have missed.
The week at Retraction Watch featured:
- Physicists flag over 50 papers on superheavy elements, leading to 3 retractions
- Guest post: A call to end the ‘impact on conclusions’ test for retraction
- Publisher to retract entire conference proceedings, ban editor who wrote most of them
- The Lancet retracts half-century-old commentary on talc for undisclosed industry ties
- Judge upholds 15-year debarment against scientist who once threatened to sue Retraction Watch
- Why don’t journalists circle back to cover retractions? A conversation with Malgorzata Iwaniec-Thompson
In case you missed the news, the Hijacked Journal Checker now has more than 400 entries. The Retraction Watch Database has over 63,000 retractions. Our list of COVID-19 retractions is up to 650, and our mass resignations list has more than 50 entries. We keep tabs on all this and more. If you value this work, please consider showing your support with a tax-deductible donation. Every dollar counts.
Here’s what was happening elsewhere (some of these items may be paywalled, metered access, or require free registration to read):
- “Conference catches illicit AI use” in hundreds of peer reviews by using watermarks.
- “Co-authorship as a traded commodity: the experiences of early career education academics.”
- “Data duplications flagged in highly cited gut-brain studies.”
- “How to build an AI scientist,” which are “changing research — institutions, funders and publishers must respond.”
- Awardees announced for program targeting changes to academic incentives. Read our Q&A about the project.
- “Seeing Is Believing? Scientific Misconduct and the Detection of Problematic Images.”
- “Dishonest behavior in biomedical journals: analysis of two cases and reflections on their detection.”
- “The Problem With Promoting ‘Gold Standard Science.’”
- University in Netherlands launches investigation into former researcher who manipulated data.
- “Devil’s advocate: the case against preprints in biomedical science.”
- “How ‘Tiny Shortcuts’ Are Poisoning Science.”
- Institutions offer “Chinese perspective on global academic journal evaluation” by releasing “global high-quality journal list for medicine, life sciences.”
- “Prompt injection in manuscripts: exploiting loopholes or crossing ethical lines?”
- “The What, Why, and How of the Paper Mill Industry for Young Scholars.”
- “Ethics tools for strengthening ethics review: A scoping Review.”
- “Even on their own terms university rankings have failed in India.”
- “Authorship order and performance from the perspective of field and academic capital: Time for a distributive ethics.”
- “STM Plants a Flag About Responsible Use of Research Content in GenAI.”
- “Do we really need to worry about hyper-prolific authors?”
- “More AI will not beat the Red Queen.”
- “Senior European journalist suspended for publishing AI-generated quotes.”
- European Research Council “sets out firm line on use of AI in peer review.”
- “The System That Decides What Science Gets Published Is Breaking Down.”
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on X or Bluesky, like us on Facebook, follow us on LinkedIn, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].
Suppose someone builds an “AI scientist”: in the case proposed, to write code, test it, and report results. Why should anyone believe it? As many posts on this site discuss, humans do not always tell the truth about their experiments, in spite of the many factors encouraging them to do so. Why would the AI scientist not merely generate fictions that resemble truthful reports of experiment sufficiently closely to pass peer review?
Ironically, the AI-produced image topping the Red Queen story shows a chess set with silver pieces and a gold queen.
Hallucinated citations:
https://www.nature.com/articles/d41586-026-00969-z
“Spokespeople for all five publishers said that they check references as part of their screening and editing process”
( Not intended to be a factual statement, but a politely framed declaration of indifference. Suitable for April 1, as it happens.)
Indeed, it is BS because the last week I yet again received review requests for two manuscripts with hallucinated references. While the other was from Elsevier’s portfolio, it seems that particularly Springer is affected. With that and APCs, they are quickly becoming the new MPDI.
They don’t even do basic copy editing. I routinely see glaring formatting errors in the reference sections of published papers. The software used is not very intelligent about references (it is rather better at handling the body) and will routinely introduce a number of errors if not overridden. This is straightforward, but you have to pay a professional a little bit if you want it done properly, so it is not done.
(As it happens, at the moment I’m reading one such from a perfectly respectable Elsevier journal. A fine article, aside from the slovenly production.)