Tag

Research Integrity

All articles tagged with #research integrity

science5 hours ago•5 min saved

Tampered Images Question the Reliability of Commercial Antibodies

A postdoctoral researcher group led by Reese Richardson and Sholto David found widespread image manipulation in marketing images for about 17,500 commercially sold antibodies from 16 companies, with roughly 7% of examined images showing edits like brightness/contrast changes, background noise removal, or copied/pasted data. These manipulations can mislead researchers about an antibody’s effectiveness, risking wasted time and money on troubleshooting experiments that may never yield reliable results.

via Ars Technica|

#antibodies #biotechnology #image-manipulation

technology13 days ago•6 min saved

Exposé: The ‘100% Human’ Medical-Research Service Built on AI and Fake Identities

An investigation finds Research Gold’s claim of 100% human-written medical research work is false: supposed PhD methodologists are AI-generated or using real researchers’ identities without consent, with AI-generated profiles and responses; at least one real researcher is seeking a takedown after learning her identity was used; the firm quoted a $1,900 price for a systematic review and appeared to respond with AI-authored messages. The report highlights broader concerns about AI’s role and potential fraud in scholarly publishing.

via 404 Media|

#academic-publishing #ai #misinformation

science-policy26 days ago•10 min saved

Fix the incentives to curb AI-driven fraud in scientific publishing

Generative AI accelerates the publish-or-perish pressure by enabling rapid fabrication of papers, with detectors often lagging behind increasingly convincing AI-made content. While disclosure norms and screening tools help, experts say the core fix is systemic: reward quality over quantity, adopt data attestations and provenance standards, and align funders, institutions, and publishers to slow AI-enabled fraud. Detection buys time, but meaningful reform requires changing the incentives that drive researchers to publish large volumes.

via C&EN|

#ai-in-science #paper-mills #peer-review

science1 month ago•11 min saved

Nepal medical student linked to fast-turnaround research network under investigation

Retraction Watch reports that Raghabendra Kumar Mahato, a second-year medical student in Nepal, runs a WhatsApp network of about 1,000 researchers producing large volumes of conference abstracts and journal papers. Several abstracts coauthored by Mahato have been accepted at major conferences such as the American Heart Association and ESMO. Critics say the setup resembles a paper mill and that authorship can be traded for presenting or publication fees, while Mahato denies selling authorship and says the aim is collaborative research; AHA and ESMO say they are reviewing the submissions.

via Retraction Watch|

#authorship #conferences #nepal

science1 month ago•4 min saved

Zurich neuropathologist cleared in misconduct probe, but seven papers flagged for errors

After reviewing 36 publications from 1996–2023, the University of Zurich cleared Adriano Aguzzi of research misconduct, finding no proven intent or negligence, though seven papers contained scientifically significant errors and should be retracted or corrected; two papers have already been retracted, sparking calls for broader accountability for co-authors. Aguzzi plans to correct the record and will continue not-for-profit research via a foundation, while experts urge improved data preservation and independent verification of key results.

via The Transmitter|

#adriano-aguzzi #pubpeer #research-integrity

science1 month ago•7 min saved

Paper mills inflate cancer research metrics by doubling citations

A large analysis of 33,159 papers across 20 high‑impact molecular-oncology journals finds papers likely produced by paper mills receive about twice as many citations as genuine papers, often citing other suspect papers and inflating journal impact factors. Using a BERT-based detector on bioscience articles, 12.3% of papers were flagged; Nature Cancer had none; patterns suggest coordinated manipulation and risks for publishers, though some false positives may exist.

via Nature|

#cancer-research #citations #impact-factor

technology2 months ago•6 min saved

Pre-submission tool flags suspect journals

A free platform called Journal Trends analyzes journal publishing patterns using OpenAlex data and the Problematic Paper Screener to help researchers identify potentially unreliable journals before submitting manuscripts, highlighting red flags such as sudden surges in papers from a single country; while spikes aren’t definitive proof of misconduct, the tool can aid due diligence and integrity investigations and inform delisting and indexing considerations like Scopus.

via Nature|

#journal-trends #openalex #problematic-paper-screener

health2 months ago•4 min saved

Retraction casts doubt on long-published autism-vaccine link study

A 16-year-old study claiming that infant hepatitis B vaccination increases autism risk has been retracted for fundamental methodological flaws, including too few autism cases (31) and overstated causality. The authors dispute the withdrawal, and the paper had previously been cited in CDC/ACIP discussions, influencing vaccine-safety reviews before the retraction was issued.

via Retraction Watch|

#autism #health #hepatitis-b

science2 months ago•5 min saved

Funding clamps, doctored data claims, and AI in science dominate Retraction Watch’s Weekend Reads

Retraction Watch’s Weekend Reads roundup highlights the White House's plan to tighten grant oversight and ban federal funding for open-access fees, sleuths’ claims that Thermo Fisher doctored data to sell antibodies, debates in China about Nature’s reputation amid AI-generated cover edits, and a broad slate of integrity and publishing stories—from peer-review issues and funding incentives to data-management reforms—plus resources like the Retraction Watch Database and the Ctrl-Z Award.

via Retraction Watch|

#biotechnology #research-integrity #science

science3 months ago•5 min saved

AI-generated fake citations flood scientific literature, study finds

A Lancet study led by Columbia University researchers finds a growing share of citations in scientific papers are fabricated or AI-generated. Analyzing 2 million papers and 97 million citations, they identified about 4,000 fabricated references across 2,800 papers, with the rate rising from 2023 to 2025 and into early 2026. The rising prevalence could skew systematic reviews and clinical guidelines and reflects broader shifts in citation culture toward checklist-style referencing. Publishers report varying levels of tooling to flag or validate citations, underscoring the need for stronger verification as AI use in research grows.

via statnews.com|

#academic-publishing #ai #fabricated-citations

technology3 months ago•11 min saved

Why Some Researchers Are Choosing to Skip Generative AI

A Nature Career feature profiles researchers who deliberately avoid generative AI, citing ethics (copyright and consent), transparency concerns, environmental costs, and a belief that relying on AI can undermine skill development; while surveys show mixed adoption with some using AI for editing or literature review, many academics argue the downsides—such as hallucinations and questionable accuracy—mean AI should be restricted in certain tasks and prompts a rethink of training and assessment in science.

via Nature|

#academia #environmental-impact #ethics

science4 months ago•6 min saved

Guest-Edited Journals Face Scrutiny Over Quality and Retractions

The BMJ group’s retraction of nearly an entire guest-edited special issue on cancer immunotherapies highlights broader worries about guest-edited publishing: rapid, high-volume special issues driven by APCs can weaken peer review, inflate researchers’ CVs, and undermine research integrity, prompting calls for tighter editorial oversight and reforms in author-pay models and funding practices.

via statnews.com|

#article-processing-charges #guest-editors #peer-review

science4 months ago•8 min saved

Dataset unmasks the price range of fake first-author slots in the paper-mill market

A dataset called BuyTheBy compiles over 18,000 advertisements from seven paper mills across multiple countries, revealing first-author slots selling for roughly $56 to $5,631 and showing some ads align with published papers. The study notes the market is evolving with AI, making enforcement difficult, but provides a starting point for publishers to curb fraud. The dataset isn’t comprehensive (omits Iran/China mills) and can’t confirm completion of purchases or final publications.

via Retraction Watch|

#academic-publishing #authorship-fraud #buytheby

science4 months ago•5 min saved

LLMs Aren’t the Problem, Cash-for-Review Fails, and Vaping Studies Reveal Flaws

Retraction Watch’s weekend digest notes that large language models aren’t the core issue in science publishing, reports that offering cash to spot errors doesn’t work, and spotlights vaping studies with numerous flaws and few retractions, while also outlining ongoing investigations and policy discussions around scientific integrity and publishing practices.

via Retraction Watch|

#llms #peer-review #research-integrity

technology6 months ago•7 min saved

AI Slop Tests the Limits of Computer Science Publishing

Nature reports that a surge of AI-generated, low-quality submissions—dubbed 'AI slop'—is flooding computer science journals and conferences, with ICML 2026 receiving over 24,000 papers and arXiv submissions up more than 50% since ChatGPT; some papers are AI-generated or contain fabrications, prompting arXiv and conference policy changes, expanded reviewer pools, and debates about moving to rolling-journal models to preserve research integrity.

via Nature|

#ai #arxiv #machine-learning