The DOI matches its Crossref record; 37 of the 37 references it cites were found in a registry or an index; 12 of the 29 sources its claims cite were checked for support.
Is the future of peer review automated?
Open accessAbstract
The rising rate of preprints and publications, combined with persistent inadequate reporting practices and problems with study design and execution, have strained the traditional peer review system. Automated screening tools could potentially enhance peer review by helping authors, journal editors, and reviewers to identify beneficial practices and common problems in preprints or submitted manuscripts. Tools can screen many papers quickly, and may be particularly helpful in assessing compliance with journal policies and with straightforward items in reporting guidelines. However, existing tools cannot understand or interpret the paper in the context of the scientific literature. Tools cannot yet determine whether the methods used are suitable to answer the research question, or whether the data support the authors’ conclusions. Editors and peer reviewers are essential for assessing journal fit and the overall quality of a paper, including the experimental design, the soundness of the study’s conclusions, potential impact and innovation. Automated screening tools cannot replace peer review, but may aid authors, reviewers, and editors in improving scientific papers. Strategies for responsible use of automated tools in peer review may include setting performance criteria for tools, transparently reporting tool performance and use, and training users to interpret reports.
Overview
CoChat Summary
TL;DR
Automated screening tools can help authors, editors, and reviewers check compliance and basic reporting-guideline items, but they cannot judge journal fit, methodological suitability, or whether data support conclusions. The paper argues tools should support, not replace, human peer review, with transparent performance reporting and user training.
Key findings
- Automated tools have the most potential in assessing compliance (ethics, consent, plagiarism, funder and journal requirements) and routine reporting-quality items such as CONSORT, PRISMA, or ARRIVE elements; broader research quality and journal fit are "best left to expert human reviewers and editors."
- Existing tools cannot understand or interpret research in the context of the scientific literature, cannot determine whether methods suit the research question, and cannot judge whether data support the authors' conclusions.
- The ScreenIT pipeline, including a growing set of automated tools, has been used to post public reports on more than 23,000 bioRxiv and medRxiv COVID-19 preprints, demonstrating feasibility of widespread screening.
- Tools can detect both problems (misreported p-values, missing ethics statements, bar graphs of continuous data, rainbow color maps, citations of retracted papers) and beneficial practices (open data, open code), and can be adapted to different study designs.
Methods
- This is a commentary/opinion piece, not an empirical study; it synthesizes prior literature and the authors' experience with automated screening.
- It describes examples of existing tools (e.g., SciScore, ODDPub, Limitation-recognizer, Barzooka, Jetfighter, Statcheck, Seek & blastn) in a table, and the ScreenIT multi-tool pipeline.
- Proposed strategies for responsible use include setting performance criteria, transparently reporting tool performance (F1, sensitivity, specificity) and use, and training users to interpret reports.
Limitations
- Stated: tools make mistakes, cannot always determine whether an item is relevant to a paper (especially with poor reporting), and reporting-quality tools may not capture methodological quality; many of these limitations may also apply to untrained human reviewers.
- Stated: integrating tools into manuscript submission systems is time consuming and expensive, and costs often exceed open-source developers' project budgets.
- Evident from the text: the paper presents no primary data, systematic review, or formal evaluation of tool performance; its claims rest on cited literature and the authors' experience, and it notes open questions (when to share reports, whether preprint reports should be public, protecting against gaming) remain unresolved.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
9 Valid · 0 Not valid · 12 Not checked · 20 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“Statistical errors are also common [7, 8].”
§ Is the future of peer review automated? > Introduction
What the sources say
- Supportedcites [7] Nuijten MB, Hartgerink CHJ, van Assen MALM, Epskamp S, Wicherts JM. The prevalence of statistical reporting errors in psychology (1985–2013). Behav Res Methods. 2016;48:1205–26. https://doi.org/10.3758/s13428-015-0664-2. source
The Results section, ‘Prevalence of inconsistencies as found by other studies,’ says reporting inconsistencies are high and that roughly one in ten results is erroneously reported. This supports the broad claim that statistical errors are common, specifically in the psychology literature studied. (LLM, conf=0.94)
- Supportedcites [8] Brown NJL, Heathers JAJ. The GRIM Test. Soc Psychol Personal Sci. 2017;8:363–9. https://doi.org/10.1177/1948550616673876. source
The indexed Abstract reports that 36 of 71 testable psychology articles appeared to contain at least one inconsistent mean. It also says the authors confirmed at least one reporting error in each of the nine cases for which they obtained data, supporting the broad claim that statistical errors are common. (LLM, conf=0.86)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Correcting errors after publication is extremely burdensome [2, 3]; hence, focusing on prevention may be more efficient.”
§ Is the future of peer review automated? > Introduction
What the sources say
- Supportedcites [2] Vorland CJ, Brown AW, Ejima K, Mayo-Wilson E, Valdez D, Allison DB. Toward fulfilling the aspirational goal of science as self-correcting: a call for editorial courage and diligence for error correction. Eur J Clin Invest. 2020;50: e13190. https://doi.org/10.1111/eci.13190. source
In “Toward an Improved Post-Publication Editorial Process” (page unavailable), the paper says error identification and correction require expertise and time, investigations can be protracted, and editorial practices can impede corrections. These passages support the characterization of post-publication correction as highly burdensome. (LLM, conf=0.85)
- Supportedcites [3] Besançon L, Bik E, Heathers J, Meyerowitz-Katz G. Correction of scientific literature: too little, too late! PLoS Biol. 2022;20: e3001572. https://doi.org/10.1371/journal.pbio.3001572. source
The Abstract says that even basic formal corrections can take months of correspondence and that authors trying to correct their own papers can face strenuous challenges. This supports the cited claim that correcting errors after publication is extremely burdensome. (LLM, conf=0.94)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Evidence that peer review substantially improves reporting, or catches errors or questionable research practices, is limited [9, 10].”
§ Is the future of peer review automated? > Introduction
What the sources say
- Partly supportedcites [9] Horbach SPJM, Halffman W. The ability of different peer review procedures to flag problematic publications. Scientometrics. 2019;118:339–73. https://doi.org/10.1007/s11192-018-2969-2. sourceExperimental — accuracy not yet measured
The Theoretical framework (Diversity and expectations in peer review; page unavailable) says very little is known about how effectively peer review detects erroneous or fraudulent research. The paper also discusses questionable research practices, but the cited passages do not establish that evidence for peer review substantially improving reporting is limited. (LLM, conf=0.86)
- Partly supportedcites [10] Blanco D, Altman D, Moher D, Boutron I, Kirkham JJ, Cobo E. Scoping review on interventions to improve adherence to reporting guidelines in health research. BMJ Open. 2019;9: e026589. https://doi.org/10.1136/bmjopen-2018-026589. sourceExperimental — accuracy not yet measured
The Discussion (page unavailable in our copy) reports only four randomized trials of interventions to improve adherence to reporting guidelines; an additional review against the guidelines had a slightly positive but nonsignificant effect, while suggesting their use to peer reviewers showed no benefit. This supports limited evidence that peer review substantially improves reporting. The paper does not assess whether peer review catches errors or questionable research practices. (LLM, conf=0.85)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Many “best practice” criteria are routinely neglected by reviewers and editors.”
§ Is the future of peer review automated? > Main text > How can automated screening help peer review?
What the sources say
- Partly supportedcites [28] Weissgerber TL, Garcia-Valencia O, Garovic VD, Milic NM, Winham SJ. Why we need to report more than “Data were analyzed by t-tests or ANOVA.” Elife. 2018. https://doi.org/10.7554/eLife.36163. sourceExperimental — accuracy not yet measured
Our copy supports the 'routinely neglected' half only at the level of the published papers: the Abstract reports that 'papers in our sample were routinely missing essential information about both types of tests' (95% of ANOVA papers lacking the information needed to determine the test type; 26.7% with no post-hoc test specified), and the Introduction notes the SAMPL guidelines 'have not been widely adopted'. The Discussion ('Moving towards a more transparent and reproducible future') says statistical reporting requirements 'are not standard practice for many fields and journals' and that editorial interventions had limited effect (Landis criteria 0%→16.4%; ARRIVE checklist 'had no effect'; a joint editorial series 'did not improve' reporting) — but those negative results are credited to Macleod 2017, Hair 2018 and Diong 2018, not to this paper's own finding. What the copy nowhere states is that reviewers and editors themselves routinely neglect best-practice criteria; it instead frames editors and reviewers as actors who 'need to work together to improve the quality of statistical reporting'. (LLM, conf=0.65)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“While preprints allow scientists to receive feedback before publishing their work in a journal, comments on preprints are uncommon [14].”
§ Is the future of peer review automated? > Introduction
What the sources say
- Supportedcites [14] Malički M, Costello J, Alperin JP, Maggio LA. Analysis of single comments left for bioRxiv preprints till september 2019. Biochem Med (Zagreb). 2021;31:20201. https://doi.org/10.11613/BM.2021.020201. source
The Discussion states that "less than ten percent of all bioRxiv preprints received public comments before the COVID-19 pandemic" and explicitly labels this "This low prevalence of scholarly public commenting", which directly supports the paraphrase that comments on preprints are uncommon. The Introduction reinforces it (despite 70% of servers allowing comments, and feedback being seen as a preprint advantage, commenting had not been systematically studied; the surge only came with COVID-19 in 2020). The only qualifier the citing sentence drops is the setting/time bound (bioRxiv, pre-September 2019), which the source itself generalises to scholarly public commenting. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
18 more claims · 4 with their cited sources checked
Create a free account to see all 41 claims and to verify the 17 references still open.
- Probably not a substantial claim. 15
- It may state two separate claims as one. 6
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Automated screening, based on natural language processing (NLP) and machine learning approaches, could reduce the burden of manual checking, streamline the peer review process, and contribute to better reporting quality18–20.”
From Text classification models for assessing the completeness of randomized controlled trial publications based on CONSORT reporting guidelines (2024) · cites this paper as [20]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Partly, this is already done [3, 4] but future LLMs could enhance these applications.”
In the citing paper
“Automated detection algorithms would save time and increase scrutiny. 11,12”
From Automated detection of over- and under-dispersion in baseline tables in randomised controlled trials. (2022) · cites this paper as [12]
Citations
Cites
The PRISMA 2020 statement: An updated guideline for reporting systematic reviews
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., et al. (2021) · PLoS Medicine · doi:10.1371/journal.pmed.1003583 · Cited by 5,148
Reporting animal research: Explanation and elaboration for the ARRIVE guidelines 2.0
Percie du Sert, N., Ahluwalia, A., Alam, S., Avey, M. T., Baker, M., Browne, W. J., Clark, A., Cuthill, I. C., Dirnagl, U., Emerson, M., et al. (2020) · PLOS Biology · doi:10.1371/journal.pbio.3000411 · Cited by 2,677 · cited 1×
“This may also include some routine aspects of the research and reporting quality assessment (e.g., compliance with elements of reporting guidelines, such as CONSORT [25], PRISMA [26], or ARRIVE [27]).”
BackgroundCONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials
Schulz, K. F., Altman, D. G., Moher, D., & for the CONSORT Group (2010) · PLoS Medicine · doi:10.1371/journal.pmed.1000251 · Cited by 2,296
Journal of Experimental Social Psychology
(1965) · Nature · doi:10.1038/2061099b0 · Cited by 1,883
Lee, C. J., Sugimoto, C. R., Zhang, G., & Cronin, B. (2013) · Journal of the American Society for Information Science and Technology · doi:10.1002/asi.22784 · Cited by 1,000 · cited 1×
“While many papers benefit from peer review, problematic papers are still published [1].”
BackgroundThe prevalence of statistical reporting errors in psychology (1985–2013)
Nuijten, M. B., Hartgerink, C. H. J., van Assen, M. A. L. M., Epskamp, S., & Wicherts, J. M. (2016) · Behavior Research Methods · doi:10.3758/s13428-015-0664-2 · Cited by 510 · cited 1×
“Statistical errors are also common [7, 8].”
Sugimoto, C. R., Ahn, Y.-Y., Smith, E., Macaluso, B., & Larivière, V. (2019) · The Lancet · doi:10.1016/s0140-6736(18)32995-7 · Cited by 305 · cited 1×
“Examples include incomplete reporting of criteria to assess risk of bias [6], ambiguous or incorrect citations [29], lack of open data or code [30], incorrect statistical calculations [31], and underreporting of ethics [32], sex as a biological variable [33], and limitations statements [34].”
BackgroundDuration and quality of the peer review process: the author’s perspective
Huisman, J., & Smits, J. (2017) · Scientometrics · doi:10.1007/s11192-017-2310-5 · Cited by 303 · cited 1×
“Whereas traditional peer review requires time and effort [35], tools can quickly screen many papers and provide individualized feedback on some of the items included in trans-parency and reporting guidelines.”
BackgroundTracking the popularity and outcomes of all bioRxiv preprints
Abdill, R. J., & Blekhman, R. (2019) · eLife · doi:10.7554/elife.45133 · Cited by 196 · cited 1×
“The growing adoption of pre-prints offers another opportunity to use automated tools to help authors to improve their papers [13].”
BackgroundThe changing forms and expectations of peer review
Horbach, S. P. J. M., & Halffman, W. (. W. (2018) · Research Integrity and Peer Review · doi:10.1186/s41073-018-0051-5 · Cited by 179 · cited 2×
“Automated screening tools may be able to address several limitations of peer review [24].”
Background
Research works printed but not linked: 1
- eLife. Jetfighter: towards figure accuracy and accessibility. 2019. https://elifesciences.org/labs/c2292989/jetfighter-towards-figure-accuracy-and-accessibility. Accessed 29 Mar 2022.
Cited by
Alqahtani, T., Badreldin, H. A., Alrashed, M., Alshaya, A. I., Alghamdi, S. S., bin Saleh, K., Alowais, S. A., Alshaya, O. A., Rahman, I., Al Yami, M. S., et al. (2023) · Research in Social and Administrative Pharmacy · doi:10.1016/j.sapharm.2023.05.016 · Cited by 565
Hosseini, M., & Horbach, S. P. J. M. (2023) · Research Integrity and Peer Review · doi:10.1186/s41073-023-00133-5 · Cited by 328
Liang, W., Zhang, Y., Cao, H., Wang, B., Ding, D., Yang, X., Vodrahalli, K., He, S., Smith, D., Yin, Y., et al. (2023) · arXiv · doi:10.48550/arxiv.2310.01783 · Cited by 243
“Despite recent attempts that explore the potential uses of such tools in areas such as automating paper screening 24 , error identification 25 , and checklist verification 26 1 , we lack large-scale empirical evidence on whether and how LLMs may be used to facilitate scientific feedback and augment…”
BackgroundMeta‐analysis prediction intervals are under reported in sport and exercise medicine
Borg, D. N., Impellizzeri, F. M., Borg, S. J., Hutchins, K. P., Stewart, I. B., Jones, T., Baguley, B. J., Orssatto, L. B. R., Bach, A. J. E., Osborne, J. O., et al. (2024) · Scandinavian Journal of Medicine and Science in Sports · doi:10.1111/sms.14603 · Cited by 99
Self-citationOpen Science 2.0: Towards a truly collaborative research ecosystem
Thibault, R. T., Amaral, O. B., Argolo, F., Bandrowski, A. E., Davidson, A. R., & Drude, N. I. (2023) · PLOS Biology · doi:10.1371/journal.pbio.3002362 · Cited by 72
Self-citation
Metadata
- Authors
- Schulz, R., Barnett, A., Bernard, R., Brown, N. J. L., Byrne, J. A., Eckmann, P., Gazda, M. A., Kilicoglu, H., Prager, E. M., Salholz-Hillel, M., ter Riet, G., Vines, T., Vorland, C. J., Zhuang, H., Bandrowski, A., & Weissgerber, T. L.
- Affiliations
- Berlin Institute of Health at Charité - Universitätsmedizin Berlin (DE)
- Queensland University of Technology (AU)
- NeuroCure Cluster of Excellence (DE)
- Linnaeus University (SE)
- The University of Sydney (AU)
- University of California San Diego (US)
- Centre National de la Recherche Scientifique (FR)
- Cohen Veterans Bioscience (US)
- Indiana University Bloomington (US)
- Syracuse University (US)
- Corresponding author
- Weissgerber, Tracey L.
- ORCID
- Venue
- BMC Research Notes
- Published
- 2022-06-11
- Volume
- 15
- Issue
- 1
- Pages
- 203-203
- ISSN
- 1756-0500 (ISSN-L)
- PMC
- PMC9188010
- OpenAlex
- W4282933389
- Cited by
- 80
- References
- 34
- Citations by year
- 2026: 10 · 2025: 20 · 2024: 27 · 2023: 17 · 2022: 6
- FWCI
- 14.65
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- BMC Research Notes (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- QUT ePrints (Queensland University of Technology) (submittedVersion, cc-by)
- Pure Amsterdam UMC (submittedVersion, cc-by)
- Data Archiving and Networked Services (DANS) (submittedVersion, other-oa)
- Data Archiving and Networked Services (DANS) (submittedVersion, cc-by)
- Diva portal (Dalarna University Library) (submittedVersion, other-oa)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- eScholarship (California Digital Library) (submittedVersion, cc-by)
- Refubium (Universitätsbibliothek der Freien Universität Berlin) (submittedVersion, cc-by)
- Universitätsbibliothek der FU Berlin Hochschulschriftenstelle u. Dokumentenserver (cc-by)
- Zenodo (CERN European Organization for Nuclear Research) (cc-by)
- Funders
- Syracuse University
- Cohen Veterans Bioscience
- University of Illinois at Urbana-Champaign
- Medical Research Council
- National Health and Medical Research Council (1184263)
- Topics
- Social Sciences › Library and Information Sciences › Academic Publishing and Open Access
- Social Sciences › Library and Information Sciences › Publishing and Scholarly Communication
- Social Sciences › Library and Information Sciences › Academic Writing and Publishing
- Keywords
- automated peer review, automated screening tool, preprints, reporting guidelines, research quality assessment, editorial practice, peer reviewers
- MeSH terms
- Research Design, Peer Review, Research, Editorial Policies, Research Report
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-04, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-04 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Is the future of peer review automated? https://app.cochat.ai/papers/10.1186/s13104-022-06080-6
- The paper
- Schulz, R., Barnett, A., Bernard, R., Brown, N. J. L., Byrne, J. A., Eckmann, P., Gazda, M. A., Kilicoglu, H., Prager, E. M., Salholz-Hillel, M., ter Riet, G., Vines, T., Vorland, C. J., Zhuang, H., Bandrowski, A., & Weissgerber, T. L. (2022). Is the future of peer review automated? BMC Research Notes. https://doi.org/10.1186/s13104-022-06080-6
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.