MétaCan
Menu
Back to cohort
Record W4415595819 · doi:10.1073/pnas.2524787122

Reply to Singer: Strike paper mills at the root

2025· article· en· W4415595819 on OpenAlexaff
Reese Richardson, Spencer S. Hong, Jennifer A. Byrne, Thomas Stoeger, Luı́s A. Nunes Amaral

Bibliographic record

VenueProceedings of the National Academy of Sciences · 2025
Typearticle
Languageen
FieldEngineering
TopicMetallurgical Processes and Thermodynamics
Canadian institutionsScience North
Fundersnot available
KeywordsEnthusiasmAction (physics)Root (linguistics)Human resourcesCall to actionSociology of scientific knowledgeScientific evidenceRoot cause

Abstract

fetched live from OpenAlex

We thank Singer for his commentary ( 1 ) on our study ( 2 ).We share his enthusiasm for fighting industrialized scientific fraud and agree that the scientific community should prioritize developing new, scalable strategies for identifying fraudulent manuscripts pre-and post-publication.We also hope that science-of-science approaches can be used to characterize factors underlying the growth of the industry, design interventions and evaluate their efficacy.We echo Singer's sentiments and prior calls to action ( 3 , 4 ).However, the rise of industrialized scientific fraud should not be thought of as inevitable.We believe that the emergent industry for scientific fraud is most likely a consequence of a hypercompetitive and highly unequal scientific enterprise, not a consequence of human nature nor any "tragedy of the commons" ( 5 ) fundamental to human scientific behavior.In the scientific enterprise we have come to inhabit, scientists compete with their peers for a scarce pool of resources and career opportunities and evaluate each other with easily gamed metrics ( 6 ).This environment was built through a series of deliberate policy decisions.Deliberate policy decisions can therefore create a fairer, more sustainable environment where aspiring scientists and professionals are not driven to engage in systematic scientific fraud.Until this change occurs, we fear that stakeholders will be stuck employing reactive, insufficient measures to which paper mills and similar organizations can easily adapt.In the meantime, scientists should ensure that new countermeasures against scientific fraud, such as those described by Singer, are carefully designed to hamper fraudulent science while not impeding genuine science ( 7 ).Paper mills thrive in a world where increased "productivity"-however misguidedly defined-has become the apparent goal of research and scholarship.Introducing additional steps that are not meaningfully enforced or that are easy to circumvent will only gift unfair competitive advantages to defectors.For instance, in response to journals now routinely employing text-similarity software to combat plagiarism in submissions, paper mills now frequently advertise access to this software to mask plagiarism before it is discovered by the journal (one such advertisement is shown in Fig. 1 ).Singer also states that journals "that cannot guarantee authenticity will shed credibility and, eventually, impact factor".We caution that this should not be taken as a foregone conclusion.The examples presented in ref. 2 suggest journals and publishers rarely face consequences (including informal consequences like reputational harm) due to breaches in quality control, even on a large scale.Indeed, paper mills exist to facilitate the manipulation of markers of success and reputation.These markers can include citation-based metrics ( 8 ) like journal impact factor ( 9 ).We encourage further empirical study of the factors influencing the frequency and severity of measurable reputational harm to journals and publishers after large-scale breaches of publication integrity.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Direct model labels (unvalidated)

Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.

Model armCategoriesStudy designConfidence
gemmano category
Domain: not available · Genre: Commentary
About the Canadian research system: no · About a Canadian topic: no
Not applicablelow
gptMetaresearchResearch integrity
Domain: Methods · Genre: Commentary
About the Canadian research system: no · About a Canadian topic: no
Not applicablehigh
models splitAgreement compares identical category sets and study designs across arms.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.001
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Bench or experimental · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.220
Threshold uncertainty score0.172

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.001
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0010.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.018
GPT teacher head0.268
Teacher spread0.250 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Labeled directly by 2 models reading the full record.

MetaresearchResearch integrity

The models disagree on parts of this classification; every voice is preserved in the section at the end of the page.

Study designNot applicable
DomainMethods
GenreCommentary

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2025
Admission routes1
Has abstractyes

Explore more

Same venueProceedings of the National Academy of SciencesSame topicMetallurgical Processes and ThermodynamicsCategoryMetaresearchFrench-language works237,207