A week of major retractions and debates over AI, authorship and research oversight shows how integrity questions are widening across science.


Research integrity is being tested on several fronts at once: large batches of retractions, questions about who gets credit for scientific work, and AI systems that may change how papers are produced and reviewed. This week’s reporting and research put those pressures into view.
Retractions and accountability
Retraction Watch’s coverage included a computing journal’s withdrawal of 13 papers, four linked to a former editor-in-chief; Bioengineered’s retraction of 119 additional papers as an investigation continues; and a year-long failure by a Springer Nature journal to correct a paper containing a fabricated reference to a Retraction Watch cofounder. A U.S. university also asked a journal to remove a professor’s name from a paper coauthored with a researcher in China.
- A computing journal retracts 13 papers, including four from a former editor-in-chief.
- Bioengineered retracts 119 more papers while its investigation continues.
- A Springer Nature journal has yet to update an article with a fake reference to a Retraction Watch cofounder.
- A university asks a journal to remove a professor’s name from a paper with a coauthor in China.
The monitoring work behind these stories is growing. The Hijacked Journal Checker now lists more than 450 entries, the Retraction Watch Database contains over 68,000 retractions, the COVID-19 retractions list is nearing 800, and the mass resignations list has more than 50 entries. Retraction Watch invites readers to support its work.
AI puts authorship and review under pressure
Debate about AI in publishing is shifting from whether researchers use these tools to how their work should be judged and credited. Researchers argue that AI-assisted article review can outperform human reviewers at their best, while journal editor Karthik Duraisamy offers an uneasy forecast: “pre-submission AI review becomes normal, then expected, then required.” A survey finds that medical journal peer reviewers are familiar with chatbots, though reported use during peer review remains limited.
Other contributions ask whether AI can generate plausible papers faster than detection and peer review can respond, and whether hallucinated references should count as research misconduct. One analysis warns that systems shaped by publishing incentives may optimize for publishability rather than truth, producing work that sounds coherent but lacks empirical grounding. Researchers are also examining disclosure standards for AI-generated images, authorship in generative AI and gaps in consensus around AI policy. An accompanying editorial cautions that agreement alone doesn’t show a policy is appropriate, effective or proportionate.
- A brief, incomplete history of modern scientific sleuths, by Retraction Watch’s cofounders.
- Researchers say AI article review can be “better than humans at their best”; an editor shares an “uncomfortable” prediction.
- AI-generated findings renew scrutiny of authorship and disclosure rules.
- An analysis of how scientific publishing incentives can shape AI outputs.
- A call to treat AI-hallucinated citations as research misconduct.
- Can detection tools and peer review keep pace with AI-generated papers?
- Disclosure and transparency standards for AI-generated images in academic publications.
- Researchers identify gaps in agreement on AI policies, while an accompanying editorial says consensus doesn’t establish a policy’s value.
- A survey finds chatbot use in medical journal peer review remains limited.
- What genomics and big data can teach us about authorship in the age of generative AI.
- How research platforms might handle the AI “slop” deluge.
Institutions, incentives and the record of science
Several stories focus on the structures around research: who can raise concerns, how institutions investigate them and what academic prestige rewards. A panel argues that libraries could help create clearer routes for reporting integrity concerns. Research on editorial mutinies considers what resignations reveal about academic status, while a separate analysis weighs transparency against ethical principles in misconduct investigations.
The journal article itself is also under scrutiny. Commentators question whether its fixed format serves current research, and an essay argues that the article’s status as science’s central currency is weakening. Other work examines conflicts of interest in scientific writing, including a study of nicotine-replacement researchers with ties to manufacturers. Disclosure notices, one critique argues, don’t by themselves shield scholarship from corporate influence.
- What editorial mutinies reveal about academic prestige.
- A panel calls for clearer channels to share research integrity concerns.
- How institutions should balance ethical principles and transparency in misconduct investigations.
- An essay argues that the journal article’s currency is collapsing; a professor asks whether the format is too static for the research it now serves.
- Why conflict-of-interest notices may not protect scientific writing from corporate influence, alongside a case study of nicotine-replacement research authors with manufacturer ties.
- An effort to quantify undetected severe misconduct in medical research.
- A 20-year analysis of retracted articles in psychiatry.
- A metascientist challenges the unified ideal of science behind the NSF’s new Office of Metascience.
Trials, trust and scientific correction
In China, a surge of controversial clinical trials is slowing after patient deaths, including a case first reported by Retraction Watch. Other coverage tracks a researcher being impersonated again and a dispute in which a Cell Press inquiry cleared an editor while a Chinese biologist continued to press the matter. A BBC podcast also looks at Retraction Watch and its Ctrl Z Award.
Elsewhere, researchers have described how they debunked “race science,” calling the effort “a triumph of scientific self-correction.” A study suggests that fossils of dinosaurs, mammals and other familiar species appear disproportionately often in high-profile journals. And a health study’s flaws have prompted a personal account asking whether one family paid a price.
- Controversial clinical trials in China slow following patient deaths, including a gene-editing trial death previously reported by Retraction Watch.
- A researcher describes being impersonated, then reports another impersonation.
- A Cell Press probe clears an editor as a Chinese biologist continues to press the issue.
- The BBC’s Unexpected Elements podcast covers Retraction Watch and the Ctrl Z Award.
- Researchers recount how they debunked “race science”.
- A study suggests fossils of familiar species are over-represented in high-profile journals; see the study preprint.
- A personal account examines the consequences of a flawed health study.
- Climate skeptics use AI to review their work and are told they are biased.
Other items include a report that the FDA is enforcing clinical trial result transparency; an article about the continued publication of papers by Putin’s sanctioned daughters in Western journals; and a discussion of the implications of a dismissal of Francesca Gino’s contract and discrimination claims against Harvard Business School.
Upcoming discussion
The 2026 Council of Science Editors Fall Virtual Symposium will feature Retraction Watch cofounder Ivan Oransky in a November 19 session, “Rethinking Incentives to Publish: Aligning Accountability Across Stakeholders.”
Support Retraction Watch with a tax-deductible contribution. Follow the organization on X, on Bluesky, on Facebook or on LinkedIn; add its RSS feed or subscribe to the daily digest. To report a retraction missing from the database, send a tip. Comments and feedback can be sent to [email protected].
Support Retraction Watch with a tax-deductible contribution. Follow the organization on X, on Bluesky, on Facebook or on LinkedIn; add its RSS feed or subscribe to the daily digest. To report a retraction missing from the database, send a tip. Comments and feedback can be sent to [email protected].
Discussion
Sign in to join the discussion.