Peer review statistics: acceptance rates, rejections, review times, and what the evidence says
143 statistics on peer review, each quoted from its primary source and verified 12 September 2026 and 20 September 2026: acceptance rates, desk rejection, reasons for rejection, review times, reviewers, reliability, bias, retractions, reproducibility, AI. Cite freely.
Fourteen numbers to quote.
How many papers are published each year, and how fast is it growing?
The pressure on peer review starts with volume. Crossref, the DOI registry used by nearly every scholarly publisher, holds records for more than twice as many journal articles published in 2025 as in 2015, and the count of indexed articles rose roughly 47% between 2016 and 2022 while the number of practicing scientists barely moved. Submissions grow faster than publications, and review invitations grow faster than either.[11][9][6]
Crossref lists 6,628,652 journal articles with a 2025 publication date, up from 3,342,205 for 2015. [11]
The number of articles indexed in Scopus and Web of Science was about 47% higher in 2022 than in 2016. [9]
Science's overall growth rate is 4.10% a year, a doubling time of 17.3 years. [10]
Since 2013, submissions have grown 6.1% a year while publications have grown 2.6% a year. [6]
Review invitations are growing 9.8% a year, twice as fast as accepted articles (4.9%). [6]
What percentage of submitted papers are accepted?
Across all journals the average acceptance rate is somewhere around 35–40%, higher in biomedicine than in the social sciences.[1] At the top of medicine the number is an order of magnitude lower, and it is falling: JAMA Pediatrics' research acceptance rate dropped from 7% in 2024 to 5% in 2025 while submissions rose.[55] The JAMA Network journals are unusual in publishing these figures every year; most journals do not, which is why the table at the bottom of this page exists.
The global average journal acceptance rate is around 35–40%, higher in biomedicine than in the social sciences. [1]
NEJM receives more than 16,000 submissions a year and publishes about 5% of original research submissions. [52]
JAMA received 11,526 manuscripts in 2025 and accepted 10.4% overall and 3.9% of research manuscripts. [53]
JAMA Internal Medicine accepted 12% of 4,178 manuscripts in 2025, and 4% of research manuscripts. [54]
JAMA Network Open received 20,233 manuscripts in 2025 and accepted 9.9%. [62]
Across nine JAMA Network journals, 2025 research acceptance rates ranged from 3.9% (JAMA) to 10% (JAMA Otolaryngology). [53][54][55][56][58][59][60][57][61]
Gut accepted 12% of submissions in 2025; the British Journal of Sports Medicine accepted 17%. [63][64]
At PLOS Medicine, about 10% of initial submissions become published articles. [65]
Elsevier journal pages that disclose a rate show Ophthalmology at 10%, Annals of Physical and Rehabilitation Medicine at 11%, the Journal of Allergy and Clinical Immunology at 20%, and Resuscitation at 28%. [66]
At the Canadian Journal of Anesthesia, 57% of 213 submissions were accepted; Canadian submissions had the highest acceptance frequency (86%). [2]
At AJR, rejection rates for the 18 countries with more than 50 submissions ranged from 22.6% to 73.4%. [3]
How many papers are rejected without peer review (desk rejection)?
Most manuscripts at selective journals never reach a reviewer. JAMA reports the share of research manuscripts it sends out for external review; in 2025 it was 23%, which means about three in four were decided by editors alone.[53] The decision is usually fast, but not always: a third of journals take more than two weeks to say no without review.[5]
JAMA sent only 23% of research manuscripts for external peer review in 2025. [53]
JAMA Internal Medicine rejected 78% of the manuscripts it received in 2017 without review; the pre-review rejection rate at Parasites & Vectors was 39%. [4]
In 2025 the median time to a first decision without peer review at JAMA Network journals ranged from 1 day (JAMA Internal Medicine) to 10 days (JAMA Surgery). [53][54][55][56][58][59][60][57][61]
One-third of journals take more than two weeks to issue a desk rejection, and one-sixth take more than four weeks. [5]
Editors screening without review are checking fit, novelty, and whether the manuscript is complete enough to be worth a reviewer's hours. The reasons are documented and predictable; see desk rejection prevention for the patterns, and the next section for the evidence.
Why are papers rejected?
In a content analysis of 151 medical-education manuscripts, the top reasons for rejection were inappropriate or incomplete statistics, overinterpretation of results, and inappropriate or suboptimal instrumentation. [12][14]
Reviewers found flaws in the methods of 85.7% of 42 manuscripts submitted to medical journals in Africa and Asia, and in the references of 69.0%. [13]
Lack of new or useful knowledge was the most common reason for rejection at AJR, accounting for 44–76% of rejections in every country. [3]
Respiratory Care's editor ranked poor study design, inadequate description of the methods, and suboptimal reporting of results among the top reasons manuscripts are not published. [14]
Only 44% of 19,123 conference abstracts were published as full papers within six years, according to the analysis summarised by Pierson. [14]
At the Canadian Journal of Anesthesia, the relationship between design, results and conclusions, and originality, drove decisions; presentation issues were rarely cited. [2]
How long does peer review take?
Two clocks run in parallel. The reviewer's clock is short: a median of 16.4 days from accepting an invitation to returning the report, and a median of five hours of actual work.[6] The author's clock is long, because finding reviewers takes more invitations every year and manuscripts wait between steps. Authors reporting their experience to SciRev waited an average of 13 weeks for a first decision, and a tenth waited more than six months.[5]
Authors reported an average first-response time of 13 weeks, ranging from 8–9 weeks in medicine and public health to 16–18 weeks in economics, mathematics and the humanities. [5]
About one-third of authors (32%) waited three months or more for a first decision, and 10% waited more than six months; rejected manuscripts took four weeks longer. [5]
Reviewers take a median of 16.4 days to complete a review after accepting the invitation (mean 19.1 days). [6]
Editors needed 2.4 review invitations to get one completed review in 2017, up from 1.9 in 2013. [6]
JAMA's median time to a first decision with peer review was 25 days in 2025; JAMA Cardiology's was 51 days. [53][58][61][62][55][56][57][54][59][60]
At JAMA Internal Medicine the median time from receipt to publication was 122 days in 2025. [54]
Elsevier's published timelines range from 1 day to a first decision at The Lancet Infectious Diseases to 126 days to a post-review decision at The Lancet Digital Health. [66]
How much work is peer review, and who does it?
Peer review is unpaid labor at a scale that is easy to miss because it is spread across millions of people. Two independent estimates put the annual total in the tens of millions to more than a hundred million hours, and both find the load concentrated on a small fraction of researchers.[7][8] The geography is skewed too: the regions that produce a growing share of papers do a shrinking share of the reviewing.[6]
Reviewers worldwide spent more than 100 million hours on peer review in 2020, the equivalent of more than 15,000 years. [7]
The unpaid reviewing time of US-based reviewers was worth over $1.5 billion in 2020; China-based reviewers over $600 million; UK-based close to $400 million. [7]
An estimated 63.4 million hours went into biomedical peer review in 2015, and 20% of researchers performed 69–94% of the reviews. [8]
Among contributing researchers, 70% spent 1% or less of their research time on peer review, while 5% spent 13% or more. [8]
On Publons, about 10% of reviewers are responsible for about 50% of peer review records. [6]
75% of journal editors say finding reviewers and getting them to accept invitations is the hardest part of their job. [6]
The USA contributes 32.9% of reviews but 25.4% of articles; China contributes 8.8% of reviews but 13.8% of articles. [6]
Emerging regions produced 29.3% of Web of Science articles but performed 18.9% of ScholarOne reviews (2013–2017), and 96.1% of editors are from established regions. [6]
A reviewer spends a median of 5 hours writing each review; after desk rejections, submissions average 2.7 reviews each. [6]
The most common reasons reviewers decline: too busy with their own research (70.6%) and the article being outside their expertise (42%). [6]
Offering reviewers $250 raised the share of invitations that produced a completed review from 42.2% to 49.8% at Critical Care Medicine, cut review time by about a day, and did not change review quality. [82]
How reliable is peer review? Do reviewers agree?
When two reviewers read the same manuscript, their recommendations agree only slightly better than chance. That is the consistent finding of the meta-analysis and of single-journal audits, and it explains why the same paper can be rejected at one journal and accepted at the next. Editors, for their part, follow unanimous reviewers and use judgment when reviewers split.[15][16]
Across 48 studies and 19,443 manuscripts, agreement between reviewers is low: mean intraclass correlation 0.34 and mean Cohen's kappa 0.17. [15]
At the Journal of General Internal Medicine, chance-corrected agreement between reviewers on rejection was kappa 0.11 across 5,881 reviews of 2,264 manuscripts. [16]
Editors rejected 88% of manuscripts when all reviewers agreed on rejection (7% of manuscripts) and 20% when all reviewers agreed the manuscript should not be rejected. [16]
When 12 already-published articles were resubmitted to the journals that had published them, only 3 of 38 editors and reviewers noticed, and 8 of the 9 that went through review were rejected. [17]
Across 16,006 manuscripts at one journal, reviewers disagreed with each other in the first round 74.8% of the time for papers later published and 85.8% for papers rejected; 6.3% of published papers had a majority recommendation to reject in the final round. [89]
How often do reviewers miss errors?
BMJ reviewers found an average of 2.58 of nine major errors deliberately inserted into a test paper; training raised this only to about 3. [18]
Reviewers who recommended accepting a fictitious manuscript identified 17.3% of its major errors; those who recommended rejection identified 39.1%. [20]
In a randomized trial at the BMJ, reviewers commented on a mean of two weaknesses in a paper containing deliberately inserted weaknesses, regardless of blinding or signing. [19]
Only 23% of 107 leading biomedical journals use specialist statistical review for all articles; 34% rarely or never do, little changed since 1998. [42]
Adding a statistical reviewer improved the quality of the final manuscript (standardized mean difference 0.58). [24]
First-round review reports on oncology trials had a median length of 276 words; 19% were constructive, 25% commented on the adequacy of conclusions, 9% on applicability and 7% on open research practices. [80]
Reviewers find roughly a quarter to a third of major methodological errors, and journals rarely add a statistician. Errors that survive review are the ones that later drive corrections, retractions and failed replications. Checking a manuscript against the relevant reporting checklist before submission is the one intervention with consistent evidence behind it.[18][42][23]
Does peer review actually improve papers?
Yes, measurably, and less than its reputation suggests. The best before-and-after study found improvement on almost every quality item, but the size of the improvement was modest; the systematic reviews conclude that the effects of editorial peer review remain largely untested.[21][22]
At Annals of Internal Medicine, 33 of 34 manuscript-quality items improved after peer review and editing; the share of items scored 3 or more rose by 7.3 percentage points from a 75% baseline. [21]
A 2002 systematic review found editorial peer review "largely untested and its effects are uncertain"; only 19 of 135 identified studies met its criteria. [22]
The Cochrane review found no clear-cut evidence that concealing reviewer or author identity affects quality (9 studies), some evidence for checklists (2), and no evidence that reviewer training helps (1). [23]
Open peer review slightly improved the quality of review reports (SMD 0.14) and lowered the odds of rejection (OR 0.56) across randomized trials. [24]
An 18-week peer-review course for doctoral students raised the editor-judged quality of their reports by a median of 0.5 points on the Review Quality Instrument. [90]
What do researchers think of peer review?
Researchers are more satisfied with peer review than the reliability data would predict, and satisfaction has risen over a decade of surveys. They also draw a sharp line around what counts: assessment by at least two researchers is peer review; a look from one editor is not.[26][25]
Nine out of ten authors believe peer review improved the last paper they published, in a 2009 survey of more than 4,000 researchers. [25]
Researcher satisfaction with peer review rose from 69% in 2009 to 75% in 2019. [26]
90% of researchers surveyed in 2019 (91% in 2009) said peer review improves the quality of published research, and 85% said there would be no control in scientific communication without it. [26]
62% of researchers found the majority of research they encountered trustworthy, but 37% had doubts about at least some research outputs they had seen in the previous week. [26]
87% of researchers count assessment by at least two researchers as peer review; only 27% count evaluation by two editorial board members, and 12% a single editor. [26]
98% of surveyed reviewers consider peer review important or extremely important, and 84.8% think institutions should explicitly recognize reviewing. [6]
Is peer review biased?
Single-blind reviewers were 1.63 times more likely to recommend papers from famous authors, 1.58 times for top universities, and 2.10 times for top companies, and bid for 22% fewer papers. [27]
The same paper was rejected by 23% of reviewers when a Nobel laureate was the visible author, 48% when anonymized, and 65% when an unknown early-career author was shown. [28]
Under open review, 40.8% of US abstracts and 22.6% of non-US abstracts were accepted at the American Heart Association meeting; blinding narrowed the gap to 33.4% versus 23.7%. [29]
Reviewers recommended a fabricated trial for publication 97.3% of the time when its result was positive and 80.0% when it showed no difference, and found more errors in the no-difference version, though the methods were identical. [30]
38% of reviews pair an editor and a reviewer from the same region, against 14% expected if reviewers were chosen in proportion to regional reviewing. [6]
How common are retractions, and what causes them?
Retractions are rare relative to output and rising in absolute terms. The 2023 spike is dominated by one publisher's mass retractions; 2024 and 2025 settled at roughly six thousand a year in the Retraction Watch Database, which we counted directly from the Crossref-hosted CSV.[33][34] Most retractions are attributed to misconduct rather than honest error, and a documented new source of unreliable papers is fabricated references produced by language models.[31][35]
Of 2,047 retracted biomedical articles, 67.4% were attributable to misconduct (fraud or suspected fraud 43.4%, duplicate publication 14.2%, plagiarism 9.8%) and 21.3% to error. [31]
The share of articles retracted for fraud rose about 10-fold between 1975 and 2012. [31]
About 2% of scientists admit to having fabricated, falsified or modified data at least once; up to 33.7% admit other questionable research practices. [32]
Retractions of research articles passed 10,000 in 2023, an annual record. [33]
The Retraction Watch Database held 66,901 retraction records on 12 September 2026; 13,226 are dated 2023, of which 9,673 are Hindawi titles, followed by 6,356 in 2024 and 5,905 in 2025. [34]
55% of the references ChatGPT-3.5 generated for literature reviews were fabricated, versus 18% for GPT-4; 43% and 24% of the real ones contained substantive errors. [35]
How reproducible are published findings?
Only 36% of 100 replications of psychology studies produced statistically significant results, and 47% of original effect sizes fell within the replication's 95% confidence interval. [36]
In 50 replicated preclinical cancer experiments, the median replication effect size was 85% smaller than the original, and 92% of replication effects were smaller than the original. [37]
How complete are published reports, and how common are statistical and reference errors?
The problems peer review is supposed to catch are common in the published record: interventions described too thinly to replicate, primary outcomes that changed between registration and publication, p-values that do not match their test statistics, and quotations that do not say what the citing paper claims.[38][39][41][43] The one bright spot is reporting guidelines: journals that endorse CONSORT publish more complete trial reports.[40]
Poor description of interventions made 40–89% of published trial reports non-replicable, and most studies had at least one primary outcome changed, introduced or omitted. [38]
Only 45.5% of 323 published trials were adequately registered, and 31% of those showed discrepancies between registered and published primary outcomes; 82.6% of assessable discrepancies favored statistically significant results. [39]
Across 50 evaluations of 16,604 randomized trials, 69 of 81 meta-analyses favored journals that endorse CONSORT for completeness of reporting. [40]
Half of published psychology papers using null-hypothesis tests contained at least one p-value inconsistent with its test statistic, and one in eight contained a gross inconsistency that may have changed the conclusion. [41]
"Standards in the use of statistics in medical research are generally low," with persistent errors reported across most medical journals. [44]
A meta-analysis of 28 studies found 25.4% of quotations in medical journal articles were inaccurate: 11.9% major and 11.5% minor errors. [43]
An audit of 2.5 million PubMed Central papers found 4,046 fabricated references in 2,810 papers; by the first seven weeks of 2026, one paper in 277 contained at least one fabricated reference, up from one in 2,828 in 2023. [68]
The fabrication rate rose more than 12-fold, from about four per 10,000 papers in 2023 to 56.9 per 10,000 in early 2026, and review articles had a 57% higher rate than other paper types. [68]
98.4% of the 2,810 papers found to contain fabricated references had received no publisher action at the time of the audit. [68]
Fabricated references were more than twice as common in medRxiv preprints (37.2 per 10,000) as in peer-reviewed PubMed Central articles (16.4 per 10,000). [70]
Of 104 medRxiv preprints with fabricated references, 12 were later published in peer-reviewed journals with the fabricated references intact. [70]
Under zero-shot, retrieval-disabled conditions, 55.0% of 300 neurocritical-care references generated by three LLMs contained an inaccuracy and 28.3% were completely fabricated; Grok-4 fabricated 50%, GPT-5.3 27% and DeepSeek-V3 8%. [93]
With live web search switched on, citation integrity rose from 96.8% to 100% for GPT-5 Thinking and from 64.6% to 98.6% for Gemini 2.5 Pro on orthopaedic guideline questions. [84]
Across 3,150 references generated by three chatbots on rotator-cuff topics, ChatGPT had the lowest reference-hallucination score and Perplexity the highest. [75]
A nursing journal that screened all 8,064 verifiable references in its 2025–2026 articles found no fabricated citations, in contrast to the wider literature. [85]
The Lancet's editors wrote that under the NIH definition of fabrication there is no debate that a fabricated reference represents research misconduct. [69]
Hallucinated citations may qualify as research misconduct under US federal regulations when they are produced by a generative-AI tool, function as data, and were not checked for accuracy. [94]
Publishers have begun building language-model tools to screen the references of submitted manuscripts, in the way plagiarism software screens text. [69]
How is AI changing peer review?
Language models are already inside the review process, mostly undeclared. Detectors classify about a fifth of ICLR reviews and an eighth of Nature Communications reviews in 2025 as AI-generated, and in a randomized experiment at ICML 2026 more than a fifth of reviewers assigned a strict prohibition reported using an LLM anyway.[45][106][107] The head-to-head studies cut both ways: on Nature-family papers an AI reviewing agent out-scored each paper's best human reviewer on criticism quality, while at an orthopaedic journal three frontier models matched desk decisions at chance level and none spotted plagiarism.[108][78] The consistent pattern is strength on checkable, rule-based tasks and weakness on judgment, with a lean toward leniency. Two further sections below cover how easily AI reviewers can be manipulated and what journals and funders now allow.[72][91][47] The full evidence review is on AI in peer review: what the evidence shows.
Between 6.5% and 16.9% of the text submitted as peer reviews to four AI conferences could have been substantially modified by large language models. [45]
GPT-4's feedback overlapped with human reviewers' points 30.85% of the time for Nature-family papers and 39.23% for ICLR, comparable to overlap between two human reviewers (28.58% and 35.25%). [46]
57.4% of researchers who tried LLM-generated feedback found it helpful or very helpful, and 82.4% found it more beneficial than feedback from at least some human reviewers. [46]
NIH prohibits its peer reviewers from using generative AI to analyze or formulate critiques of grant applications (notice issued 23 June 2023). [47]
About 20% of ICLR reviews and 12% of Nature Communications reviews in 2025 were classified as AI-generated by a detector trained on historical reviews (preprint). [106]
In a randomized experiment at ICML 2026, the LLM-use policy assigned to reviewers had near-zero effects on decisions and scores, and 22.5% of reviewers under a prohibition still reported using an LLM (preprint). [107]
Rated by 45 domain scientists on 82 Nature-family papers, a GPT-5.2 reviewing agent scored above each paper's top-rated human reviewer (60.0% vs 48.2%), but AI reviewers overlapped with each other far more than humans did (21% vs 3%) (preprint). [108]
AI review scores can be raised simply by having a language model rewrite the paper, and AI reviewers show a 'hivemind' of excessive agreement (preprint). [109]
Evaluated on ICLR 2026 and Nature Communications submissions, current LLMs generate detailed reviews but show overly positive recommendations, generic criticism and uneven evidence grounding (preprint). [110]
An author-facing LLM system covered 44.9% of the issues historical reviewers had raised on ICLR 2026 papers with a single pass, and up to 84.9% with deduplication and refill at 5.2 times the tokens (preprint). [111]
Comparing 398 human reports on 119 BMJ Open articles with ChatGPT reviews, human reviewers addressed interpretation, originality and applicability, while AI reviews covered more sections but focused on routine or structural elements. [71]
In a blinded cardiology-journal study, LLM reviews agreed with the final editorial decision 67.5% of the time versus 71.9% for the human consensus; AI correctly recommended rejection for 60% of rejected manuscripts, humans for 56%. [87]
At a sports-medicine journal a local LLM recommended revision for 61.1% of manuscripts, 72.7% of which were ultimately rejected or cascaded; when it recommended rejection or cascade, 90.5% received that decision. [81]
Three frontier LLMs matched an orthopaedic journal's desk-review decisions 59–63% of the time with near-zero kappa, systematically over-rejected accepted manuscripts, and none identified plagiarism or dual submission. [78]
The same LLMs wrote decision letters rated higher than the journal's own (Claude 4.46 vs 4.01 on a 5-point scale), a dissociation the authors describe as reproducing the surface features of peer review rather than its reasoning. [78]
Across 720 simulated radiology reviews by eight LLMs, no model ever recommended rejection, and prompt strictness shifted the severity of decisions. [77]
Human reviewers rejected 73.33% of 300 ophthalmology manuscripts; ChatGPT and Gemini rejected 2.00% each and omitted line numbers and references. [100]
ChatGPT accepted 95–98% of hand-surgery manuscripts that a higher-impact journal had rejected, yet its reviews scored higher on a review-quality instrument than the human reviews. [101]
In risk-of-bias assessment of orthopaedic trials, two AI models agreed with each other 91% of the time but deviated from the expert in 30–38.5% of interpretive domains, while matching perfectly on rule-based items. [72]
On 763 biotechnology preprints and 12 grant proposals, AI reviewers were more lenient than humans and AI detectors could not reliably identify AI-written review text. [105]
Five open-source LLMs placed transplantation papers in the correct journal quartile 35% of the time at best and showed no affiliation bias. [97]
Constrained by a rule-based prompt, an LLM chose the right statistical test for 99.3% of 148 items and reproduced chi-square values in 94.8% of tests, but anchored on authors' reported values. [91]
Giving an LLM the structured PRISMA 2020 checklist raised adherence-checking accuracy to 78.7–79.7% from 45.2% with the manuscript alone; the best model reached 95.1% sensitivity but 49.3% specificity. [79]
Asked to name expert reviewers without a diversity prompt, GPT-4o returned 68% men and 95.3% scientists from high-income countries; with the prompt, 51% women and 42.3% from high-income countries. [98]
PeerReviewAI is built for authors, not reviewers: a pre-submission review of your own manuscript, with every reference checked against PubMed and Crossref and the reporting checklist audited item by item, so the errors in the sections above are found before an editor sees them. It never issues an accept or reject verdict. See the full sample review.
Can AI reviewers be manipulated?
Yes, and easily, at least for the models tested. Hidden nudges flipped four commercial models' recommendations most of the time, invisible white-text instructions turned incoherent abstracts into favourable decisions for two of four models, and a language model rewriting a paper can raise the scores AI reviewers give it. The countermeasures are also documented: one model resisted every injection, early-2026 frontier models penalised injected text, and OCR preprocessing removed the effect entirely.[88][73][92][109]
Hidden nudges inserted into manuscripts flipped the recommendation of four commercial AI models 84.4% of the time, and warning the models about nudges barely helped (76.8%). [88]
Invisible white text in incoherent trial abstracts produced favourable LLM decisions 61% of the time versus 0% in controls; ChatGPT and Gemini were fooled 93% of the time, Claude never. [73]
Invisible text injection inflated older models' quality scores by 0.82 points, while early-2026 frontier models penalised it, and OCR preprocessing removed the effect entirely. [92]
What do journals and funders allow, and what do they require?
Most top medical journals now say something about AI in peer review, and most of what they say is a prohibition grounded in confidentiality. Author-side policies are near universal but shallow: disclosure is required almost everywhere, yet fewer than half of leading orthopaedic journals address AI-generated references or peer review, and across 5.2 million papers only about 0.1% disclosed AI use at all.[102][96][86][47] The rules that recur in every source: a human is accountable, nothing confidential goes into a third-party tool, and AI must not write the review or make the decision.[76][83][95]
Of the top 100 medical journals, 78 give guidance on AI in peer review; 59% of those prohibit it outright, 91% prohibit uploading manuscript content to AI, and 96% cite confidentiality as the reason. [102]
Across 5,114 journals and 5.2 million papers, 70% of journals had AI policies, AI-assisted writing rose regardless, and only about 0.1% of papers since 2023 disclosed AI use. [86]
Among 32 leading orthopaedic and biomedical journals, 91% had an AI policy for authors and 97% of those required disclosure, but only 45% addressed peer review and 48% AI-generated references. [96]
94 of the top 100 orthopaedics and sports-medicine journals reference AI in their policies, all prohibiting AI authorship and requiring disclosure; only 1% reference an AI-specific reporting guideline. [103]
18 of 25 vascular-surgery journals had explicit generative-AI policies; 76% required disclosure but only 56% said where. [104]
A Delphi consensus of 53 anaesthesia and pain-medicine editors-in-chief holds that LLMs must not generate data, references, conclusions or manuscripts, nor be used for editorial decisions or peer-review reports. [76]
JAMA and the JAMA Network journals extended their AI policy to peer reviewers in 2023, with reminders of the confidential nature of submitted manuscripts. [99]
Academic Medicine's editors: an AI tool cannot assume professional accountability, and peer reviewers, like authors, must be human. [83]
Reviewers should not upload the manuscript or even their own review sheets to AI tools, because deletion cannot be verified; authors are usually permitted to upload their own manuscripts for language editing. [95]
A scoping review of 189 studies found current AI systems lack the domain reasoning and ethical judgment for autonomous evaluation, with confidentiality breaches from third-party tools a key risk. [74]
ICMJE: reviewers must ask the journal's permission before using AI to help with a review, and confidentiality may prohibit uploading the manuscript to any AI technology where confidentiality cannot be assured. [112]
ICMJE: journals should require authors to disclose AI-assisted technologies at submission; chatbots cannot be authors; humans are responsible for the submitted material and must check AI output. [112]
COPE Forum advice, updated March 2026: AI-generated reviewer comments should not be used, reviewers should not upload manuscripts to AI tools, and the reviewer stays accountable for any AI suggestion included in a report. [113]
In the COPE case behind that advice, editors measured AI-content rates above 80% in suspect reviews that raised problems the manuscript did not contain. [113]
COPE's position: AI tools cannot be authors because they cannot take responsibility; authors must disclose how and which tool was used and are fully responsible for the content. [114]
Where do manuscripts go, and where did peer review come from?
Acceptance rates at the top journal in each medical field (2025–2026)
We took the Scimago Journal Rank (2024 edition), selected in each medicine category the highest-ranked journal whose title names the field, excluding review-only titles, supplements and bulletins, applied a clinician's overrides where the ranking produced a mismatch, and then looked for a published acceptance rate on each journal's own site or in its annual editorial. The result: only 17 of 49 publish one. Elsevier's Journal Insights pages publish decision timelines for most titles and an acceptance rate for a few; BMJ journals publish both; the JAMA Network publishes everything; Wiley, Nature Portfolio and several society sites publish nothing we could capture.
How these numbers were verified
Every statistic on this page was traced to its primary source and quoted from that source's own text: the authors' abstract or full text for journal articles (records retrieved through Europe PMC, PubMed and Crossref; full texts of the 20 September 2026 batch are on file, with page numbers recorded in the ledger), the report PDF for industry and survey reports, the journal's own site or annual editorial for acceptance rates, and a direct query of the Crossref API and the Crossref-hosted Retraction Watch CSV for the counts we computed ourselves. Figures that could only be found in secondary roundups, or behind paywalls we could not read, were left out. The verification ledger records the verbatim sentence, DOI or URL, and capture date for each entry.
Last verified 20 September 2026 (AI in peer review, fabricated references and process entries) and 12 September 2026 (all other entries). Journal acceptance rates and vendor prices change; the capture date is stated wherever a figure is time-sensitive. Corrections: use the contact form and cite the statistic ID shown in the ledger.
Questions, answered.
Don't see yours? Email us — we read every one.
References and sources
- [1]Bjórk Bo-Christer Acceptance rates of scholarly peer-reviewed journals: A literature survey. El Profesional de la Información 2019. doi:10.3145/epi.2019.jul.07peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [2]Turcotte C, Drolet P, Girard M. Study design, originality and overall consistency influence acceptance or rejection of manuscripts submitted to the Journal. Canadian journal of anaesthesia = Journal canadien d'anesthesie 2004. doi:10.1007/bf03018396 PubMed 15197116.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [3]Ehara S, Takahashi K. Reasons for rejection of manuscripts submitted to AJR by international authors. AJR. American journal of roentgenology 2007. doi:10.2214/ajr.06.0448 PubMed 17242215.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [4]Dantas-Torres F. Top 10 reasons your manuscript may be rejected without review. Parasites & vectors 2022. doi:10.1186/s13071-022-05543-w PubMed 36357928.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [5]Huisman J, Smits J. Duration and quality of the peer review process: the author's perspective. Scientometrics 2017. doi:10.1007/s11192-017-2310-5 PubMed 29056794.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [6]Publons. 2018 Global State of Peer Review. Clarivate Analytics; 2018. publons.com/static/Publons-Global-State-Of-Peer-Review-2018.pdfindustry report (PDF on file) · verified 2026-09-12
- [7]Aczel B, Szaszi B, Holcombe AO. A billion-dollar donation: estimating the cost of researchers' time spent on peer review. Research integrity and peer review 2021. doi:10.1186/s41073-021-00118-2 PubMed 34776003.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [8]Kovanis M et al. The Global Burden of Journal Peer Review in the Biomedical Literature: Strong Imbalance in the Collective Enterprise. 2016. doi:10.1371/journal.pone.0166387 PubMed 27832157.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [9]Hanson et al. Author response for "The strain on scientific publishing". 2024. doi:10.1162/qss_a_00327/v2/response1peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [10]Bornmann Lutz, Haunschild Robin, Mutz Rüdiger Growth rates of modern science: a latent piecewise growth curve approach to model publication numbers from established and new literature databases. Humanities and Social Sciences Communications 2021. doi:10.1057/s41599-021-00903-wpeer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [11]Crossref REST API, count of works of type journal-article by publication year. Query run 12 September 2026. api.crossref.org/works?filter=type:journal-article,from-pub-date:2025-01-01,until-pub-date:2025-12-31&rows=0database query (reproducible) · verified 2026-09-12
- [12]Bordage G. Reasons reviewers reject and accept manuscripts: the strengths and weaknesses in medical education reports. Academic medicine : journal of the Association of American Medical Colleges 2001. doi:10.1097/00001888-200109000-00010 PubMed 11553504.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [13]Ezeala C, Nweke I, Ezeala M. Common errors in manuscripts submitted to medical science journals. Annals of medical and health sciences research 2013. doi:10.4103/2141-9248.117957 PubMed 24116317.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [14]Pierson DJ. The top 10 reasons why manuscripts are not accepted for publication. Respiratory care 2004. pubmed.ncbi.nlm.nih.gov/15447812/ PubMed 15447812.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [15]Bornmann L, Mutz R, Daniel HD. A reliability-generalization study of journal peer reviews: a multilevel meta-analysis of inter-rater reliability and its determinants. PloS one 2010. doi:10.1371/journal.pone.0014331 PubMed 21179459.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [16]Kravitz RL et al. Editorial peer reviewers' recommendations at a general medical journal: are they reliable and do editors care?. PloS one 2010. doi:10.1371/journal.pone.0010072 PubMed 20386704.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [17]Peters Douglas P., Ceci Stephen J. Peer-review practices of psychological journals: The fate of published articles, submitted again. Behavioral and Brain Sciences 1982. doi:10.1017/s0140525x00011183peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [18]Schroter S et al. What errors do peer reviewers detect, and does training improve their ability to detect them?. Journal of the Royal Society of Medicine 2008. doi:10.1258/jrsm.2008.080062 PubMed 18840867.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [19]Godlee F, Gale CR, Martyn CN. Effect on the quality of peer review of blinding reviewers and asking them to sign their reports: a randomized controlled trial. JAMA 1998. doi:10.1001/jama.280.3.237 PubMed 9676667.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [20]Baxt WG et al. Who reviews the reviewers? Feasibility of using a fictitious manuscript to evaluate peer reviewer performance. Annals of emergency medicine 1998. doi:10.1016/s0196-0644(98)70006-x PubMed 9737492.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [21]Goodman SN et al. Manuscript quality before and after peer review and editing at Annals of Internal Medicine. Annals of internal medicine 1994. doi:10.7326/0003-4819-121-1-199407010-00003 PubMed 8198342.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [22]Jefferson T et al. Effects of editorial peer review: a systematic review. JAMA 2002. doi:10.1001/jama.287.21.2784 PubMed 12038911.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [23]Jefferson T et al. Editorial peer review for improving the quality of reports of biomedical studies. The Cochrane database of systematic reviews 2007. doi:10.1002/14651858.mr000016.pub3 PubMed 17443635.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [24]Bruce R et al. Impact of interventions to improve the quality of peer review of biomedical journals: a systematic review and meta-analysis. BMC medicine 2016. doi:10.1186/s12916-016-0631-5 PubMed 27287500.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [25]Mulligan Adrian, Hall Louise, Raphael Ellen Peer review in a changing world: An international study measuring the attitudes of researchers. Journal of the American Society for Information Science and Technology 2012. doi:10.1002/asi.22798peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [26]Elsevier and Sense about Science. Quality, trust & peer review: researchers' perspectives 10 years on. September 2019. senseaboutscience.org/wp-content/uploads/2019/09/Quality-trust-peer-review.pdfsurvey report (PDF on file) · verified 2026-09-12
- [27]Tomkins A, Zhang M, Heavlin WD. Reviewer bias in single- versus double-blind peer review. Proceedings of the National Academy of Sciences of the United States of America 2017. doi:10.1073/pnas.1707323114 PubMed 29138317.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [28]Huber J et al. Nobel and novice: Author prominence affects peer review. Proceedings of the National Academy of Sciences of the United States of America 2022. doi:10.1073/pnas.2205779119 PubMed 36194633.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [29]Ross JS et al. Effect of blinded peer review on abstract acceptance. JAMA 2006. doi:10.1001/jama.295.14.1675 PubMed 16609089.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [30]Emerson GB et al. Testing for the presence of positive-outcome bias in peer review: a randomized controlled trial. Archives of internal medicine 2010. doi:10.1001/archinternmed.2010.406 PubMed 21098355.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [31]Fang FC, Steen RG, Casadevall A. Misconduct accounts for the majority of retracted scientific publications. Proceedings of the National Academy of Sciences of the United States of America 2012. doi:10.1073/pnas.1212247109 PubMed 23027971.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [32]Fanelli D. How many scientists fabricate and falsify research? A systematic review and meta-analysis of survey data. PloS one 2009. doi:10.1371/journal.pone.0005738 PubMed 19478950.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [33]Van Noorden R. More than 10,000 research papers were retracted in 2023 - a new record. Nature 2023. doi:10.1038/d41586-023-03974-8 PubMed 38087103.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [34]Retraction Watch Database (Crossref). CSV downloaded 12 September 2026; yearly counts computed by PeerReviewAI from the RetractionDate field, records with RetractionNature = Retraction. gitlab.com/crossref/retraction-watch-datadatabase download (reproducible; CSV on file) · verified 2026-09-12
- [35]Walters WH, Wilder EI. Fabrication and errors in the bibliographic citations generated by ChatGPT. Scientific reports 2023. doi:10.1038/s41598-023-41032-5 PubMed 37679503.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [36]Open Science Collaboration. PSYCHOLOGY. Estimating the reproducibility of psychological science. 2015. doi:10.1126/science.aac4716 PubMed 26315443.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [37]Errington TM et al. Investigating the replicability of preclinical cancer biology. eLife 2021. doi:10.7554/elife.71601 PubMed 34874005.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [38]Glasziou P et al. Reducing waste from incomplete or unusable reports of biomedical research. Lancet (London, England) 2014. doi:10.1016/s0140-6736(13)62228-x PubMed 24411647.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [39]Mathieu S et al. Comparison of registered and published primary outcomes in randomized controlled trials. 2009. doi:10.1001/jama.2009.1242 PubMed 19724045.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [40]Turner L et al. Does use of the CONSORT Statement impact the completeness of reporting of randomised controlled trials published in medical journals? A Cochrane review. Systematic reviews 2012. doi:10.1186/2046-4053-1-60 PubMed 23194585.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [41]Nuijten MB et al. The prevalence of statistical reporting errors in psychology (1985-2013). Behavior research methods 2016. doi:10.3758/s13428-015-0664-2 PubMed 26497820.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [42]Hardwicke TE, Goodman SN. How often do leading biomedical journals use statistical experts to evaluate statistical methods? The results of a survey. PloS one 2020. doi:10.1371/journal.pone.0239598 PubMed 33002031.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [43]Jergas H, Baethge C. Quotation accuracy in medical journal articles-a systematic review and meta-analysis. PeerJ 2015. doi:10.7717/peerj.1364 PubMed 26528420.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [44]Strasak AM et al. Statistical errors in medical research - a review of common pitfalls. Swiss medical weekly 2007. doi:10.4414/smw.2007.11587 PubMed 17299669.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [45]Liang W, Izzo Z, Zhang Y, et al. Monitoring AI-modified content at scale: a case study on the impact of ChatGPT on AI conference peer reviews. arXiv:2403.07183 (v4, 19 May 2026). arxiv.org/abs/2403.07183preprint (arXiv abstract captured) · verified 2026-09-12
- [46]Liang W, Zhang Y, Cao H, et al. Can large language models provide useful feedback on research papers? A large-scale empirical analysis. NEJM AI 2024. doi:10.1056/AIoa2400196 (arXiv:2310.01783). doi.org/10.1056/aioa2400196peer-reviewed article (Crossref record; arXiv abstract captured) · verified 2026-09-12
- [47]National Institutes of Health. NOT-OD-23-149: The Use of Generative Artificial Intelligence Technologies is Prohibited for the NIH Peer Review Process. 23 June 2023. grants.nih.gov/grants/guide/notice-files/NOT-OD-23-149.htmlpolicy notice · verified 2026-09-12
- [48]Calcagno V et al. Flows of research manuscripts among scientific journals reveal hidden submission patterns. Science (New York, N.Y.) 2012. doi:10.1126/science.1227833 PubMed 23065906.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [49]Ross-Hellauer T. What is open peer review? A systematic review. 2017. doi:10.12688/f1000research.11369.2peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [50]Burnham JC. The evolution of editorial peer review. JAMA 1990. doi:10.1001/jama.1990.03440100023003 PubMed 2406470.peer-reviewed article (Europe PMC / Crossref record) · verified 2026-09-12
- [51]Rennie D. Editorial peer review: its development and rationale. In: Peer Review in Health Sciences (2003 PDF on file). book chapter (PDF on file) · verified 2026-09-12
- [52]NEJM. About NEJM; Publication Process (NEJM Media Center). Pages captured 12 September 2026. www.nejm.org/about-nejm/about-nejmjournal website · verified 2026-09-12
- [53]Bibbins-Domingo K, Christiansen SL, Curfman G, et al. 2025 Year in Review and Vision for the Path Ahead. JAMA. Published online 13 April 2026. jamanetwork.com/journals/jama/fullarticle/2847844journal editorial · verified 2026-09-12
- [54]Inouye SK. JAMA Internal Medicine—The Year in Review, 2025. JAMA Intern Med. Published online 16 March 2026. jamanetwork.com/journals/jamainternalmedicine/fullarticle/2846285journal editorial · verified 2026-09-12
- [55]Christakis DA. JAMA Pediatrics—The Year in Review, 2025. JAMA Pediatr. Published online 16 March 2026. jamanetwork.com/journals/jamapediatrics/fullarticle/2845861journal editorial · verified 2026-09-12
- [56]Kibbe MR. JAMA Surgery—The Year in Review, 2025. JAMA Surg. Published online 18 March 2026. jamanetwork.com/journals/jamasurgery/fullarticle/2846802journal editorial · verified 2026-09-12
- [57]Shinkai K. JAMA Dermatology—The Year in Review, 2025. JAMA Dermatol. Published online 18 March 2026. jamanetwork.com/journals/jamadermatology/fullarticle/2845993journal editorial · verified 2026-09-12
- [58]Bonow RO, Casadei B. JAMA Cardiology Year in Review, 2025. JAMA Cardiol. Published online 18 March 2026. jamanetwork.com/journals/jamacardiology/fullarticle/2846694journal editorial · verified 2026-09-12
- [59]Josephson SA. JAMA Neurology Year in Review, 2025. JAMA Neurol. Published online 16 March 2026. jamanetwork.com/journals/jamaneurology/fullarticle/2846475journal editorial · verified 2026-09-12
- [60]Öngür D. The Year in Review, 2025. JAMA Psychiatry. Published online 18 March 2026. jamanetwork.com/journals/jamapsychiatry/fullarticle/2846481journal editorial · verified 2026-09-12
- [61]Piccirillo JF. JAMA Otolaryngology–Head & Neck Surgery—The Year in Review, 2025. Published online 19 March 2026. jamanetwork.com/journals/jamaotolaryngology/fullarticle/2846035journal editorial · verified 2026-09-12
- [62]Fihn SD, Jacobs EA, Kim HS, Tarini BA, Perencevich EN. JAMA Network Open—The Year in Review, 2025. Published online 24 March 2026. jamanetwork.com/journals/jamanetworkopen/fullarticle/2846931journal editorial · verified 2026-09-12
- [63]Gut (BMJ). About: Journal Statistics 2025. Captured 12 September 2026. gut.bmj.com/pages/about/journal website · verified 2026-09-12
- [64]British Journal of Sports Medicine (BMJ). About: Journal Statistics 2025. Captured 12 September 2026. bjsm.bmj.com/pages/about/journal website · verified 2026-09-12
- [65]PLOS Medicine. Journal Information. Captured 12 September 2026. journals.plos.org/plosmedicine/s/journal-informationjournal website · verified 2026-09-12
- [66]Elsevier ScienceDirect Journal Insights pages for the journals named (publishing timeline and, where shown, acceptance rate). Captured 12 September 2026. www.sciencedirect.com/journal/ophthalmology/about/insightspublisher website · verified 2026-09-12
- [67]Springer Nature journal pages for European Journal of Epidemiology, Journal of Hematology & Oncology, Genome Medicine, BMC Medicine (Journal metrics, 2025). Captured 12 September 2026. link.springer.com/journal/10654publisher website · verified 2026-09-12
- [68]Topaz M et al. Fabricated citations: an audit across 2·5 million biomedical papers. Lancet (London, England) 2026. doi:10.1016/s0140-6736(26)00603-3 PubMed 42107362.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20 · Topaz et al., Lancet correspondence; audit of PubMed Central OA subset 2023-01-01 to 2026-02-18
- [69]Bauchner H, Rivara FP. Fabricated references: a new threat to editorial integrity. Lancet (London, England) 2026. doi:10.1016/s0140-6736(26)00798-1 PubMed 42107358.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20 · Bauchner & Rivara comment on the audit
- [70]Topaz M et al. Fabricated references are twice as common in medRxiv preprints as in peer-reviewed articles. Journal of internal medicine 2026. doi:10.1111/joim.70139 PubMed 42444601.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20
- [71]Rajakumar HK et al. Peer review in the age of artificial intelligence: a comparative study of human and AI-generated review reports. Postgraduate medical journal 2026. doi:10.1093/postmj/qgag005 PubMed 41591890.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [72]Lettner J et al. A new peer reviewer? Comparing AI with human performance in randomized controlled trial risk-of-bias assessment. Advances in clinical and experimental medicine : official organ Wroclaw Medical University 2026. doi:10.17219/acem/216070 PubMed 41953984.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [73]De Cassai A et al. Prompt injection compromises large language model-based peer review: evidence from randomised controlled trial abstracts from anaesthesia journals. British journal of anaesthesia 2026. doi:10.1016/j.bja.2026.06.041 PubMed 42580931.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20 · pre-registered OSF rxzjv
- [74]Nabavi A et al. Artificial intelligence in scholarly peer review: a scoping review of applications, risks, and governance challenges. International journal of medical informatics 2026. doi:10.1016/j.ijmedinf.2026.106418 PubMed 41950628.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [75]Özbek İC, Bağcıer F. Reference Hallucination in AI-Assisted Academic Writing: A Comparative Analysis of ChatGPT, Gemini, and Perplexity in Rotator Cuff Literature. Indian journal of orthopaedics 2026. doi:10.1007/s43465-026-01807-0 PubMed 42558550.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [76]De Cassai A et al. Responsible use of large language models in manuscript authorship, peer review, and editorial processes: a Delphi consensus among editors-in-chief of anaesthesia and pain medicine journals (RULE-AP). British journal of anaesthesia 2026. doi:10.1016/j.bja.2026.01.029 PubMed 41748337.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [77]Erturk SM, Durmaz M. Large Language Models as Peer Reviewers: Prompt Sensitivity and Model-Dependent Reproducibility. Academic radiology 2026. doi:10.1016/j.acra.2026.04.046 PubMed 42624573.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [78]Lupon E et al. Can large language models provide high-quality desk review decisions in an orthopaedic surgery journal? A concordance study comparing three AI models to human editorial decisions. Orthopaedics & traumatology, surgery & research : OTSR 2026. doi:10.1016/j.otsr.2026.104801 PubMed 42498040.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [79]Kataoka Y et al. Large language models for automated PRISMA 2020 adherence checking. International journal of medical informatics 2026. doi:10.1016/j.ijmedinf.2026.106561 PubMed 42364467.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [80]Logullo P et al. Peer review reports of randomized controlled trials in oncology can be short and superficial. Journal of clinical epidemiology 2025. doi:10.1016/j.jclinepi.2025.111893 PubMed 40602628.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [81]Patel R et al. Artificial Intelligence Cannot Replace Peer Reviewers but May Help Editors Triage: A Comparative Analysis of a Large Language Model and Human Reviewer Recommendations at the <i>American Journal of Sports Medicine</i>. The American journal of sports medicine 2026. doi:10.1177/03635465261463006 PubMed 42438216.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [82]Cotton CS et al. Effect of Monetary Incentives on Peer Review Acceptance and Completion: A Quasi-Randomized Interventional Trial. Critical care medicine 2025. doi:10.1097/ccm.0000000000006637 PubMed 40047491.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [83]Fischer K et al. Artificial intelligence tools in scholarly publishing: guidance for peer reviewers. Academic medicine : journal of the Association of American Medical Colleges 2026. doi:10.1093/acamed/wvaf103 PubMed 41806860.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20
- [84]Şengül HB et al. Deep research capabilities in GPT-5 thinking and Gemini 2.5 Pro improve citation integrity and concordance with American Academy of Orthopaedic Surgeons anterior cruciate ligament and rotator cuff guidelines. Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA 2026. doi:10.1002/ksa.70315 PubMed 41649196.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [85]Moons P. Fabricated references due to artificial intelligence hallucination in the European Journal of Cardiovascular Nursing: also here nurses are the most trustable profession. European journal of cardiovascular nursing 2026. doi:10.1093/eurjcn/zvag148 PubMed 42334371.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20
- [86]He Y, Bu Y. Academic journals' AI policies fail to curb the surge in AI-assisted academic writing. Proceedings of the National Academy of Sciences of the United States of America 2026. doi:10.1073/pnas.2526734123 PubMed 41734077.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [87]Zancanaro E et al. AI as a peer reviewer: a blinded comparative study of LLM-generated and human reviews in a cardiology journal. European heart journal. Imaging methods and practice 2026. doi:10.1093/ehjimp/qyag097 PubMed 42317395.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [88]Tomazini BM et al. Susceptibility of large language models to hidden nudge injection during simulated medical peer review: a quasi-experimental study. Research integrity and peer review 2026. doi:10.1186/s41073-026-00225-y PubMed 42237194.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [89]De Las Cuevas C. Reviewer recommendations and editorial outcomes in peer review: a longitudinal analysis of agreement and disagreement across review rounds. Research integrity and peer review 2026. doi:10.1186/s41073-026-00200-7 PubMed 42260682.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [90]Rohmann JL et al. Effectiveness of the researcher-led "Peerspectives" peer review training course on review quality, knowledge, and skills among doctoral students in the biomedical sciences: a pre-post study. Research integrity and peer review 2026. doi:10.1186/s41073-026-00220-3 PubMed 42210385.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [91]Dong M et al. Potential and limitations of a large language model in the statistical review of comparative categorical data: an exploratory study of structured prompt-guided approach. Research integrity and peer review 2026. doi:10.1186/s41073-026-00231-0 PubMed 42387591.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [92]Jarrett P et al. Frontier Language Models and Optical Character Recognition Preprocessing Against Invisible Text Injection in AI Peer Review. JAMA network open 2026. doi:10.1001/jamanetworkopen.2026.17356 PubMed 42258215.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20
- [93]Seifi A, Seyfi A. Hallucination Rate of Peer-Reviewed Citations Generated by Large Language Models in Neurocritical Care. Critical care explorations 2026. doi:10.1097/cce.0000000000001474 PubMed 42640622.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [94]Resnik DB, Hosseini M. Hallucinated citations produced by generative artificial intelligence may constitute research misconduct when citations function as data in scholarly papers. Accountability in research 2026. doi:10.1080/08989621.2026.2645390 PubMed 41833014.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [95]Matsubara S. Zero Risk Is Impossible: Confidentiality and Artificial Intelligence Use in Peer Review. JMA journal 2026. doi:10.31662/jmaj.2026-0060 PubMed 42577012.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20
- [96]Kelly N, Neuner J, Sabharwal S. Variability in Journal and Publisher Policies on Artificial Intelligence Use in Manuscript Preparation: An Orthopaedic Perspective. JB & JS open access 2026. doi:10.2106/jbjs.oa.26.00221 PubMed 42698649.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [97]Shen SM et al. Evaluation of Large Language Models for Peer Review in Transplantation Research: Algorithm Validation Study. JMIR AI 2026. doi:10.2196/84322 PubMed 41672474.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [98]Teixeira AL. AI in peer review: can artificial intelligence be an ally in reducing gender and geographical gaps in peer review? A randomized trial. Research integrity and peer review 2025. doi:10.1186/s41073-025-00182-y PubMed 41139793.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [99]Flanagin A, Kendall-Taylor J, Bibbins-Domingo K. Guidance for Authors, Peer Reviewers, and Editors on Use of AI, Language Models, and Chatbots. JAMA 2023. doi:10.1001/jama.2023.12500 PubMed 37498593.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20
- [100]Moshirfar M et al. Current capabilities of large language models as peer reviewers for manuscripts submitted to ophthalmology-related journals. Medicine 2026. doi:10.1097/md.0000000000048147 PubMed 41894292.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [101]Marrella D et al. Comparing AI-generated and human peer reviews: A study on 11 articles. Hand surgery & rehabilitation 2025. doi:10.1016/j.hansur.2025.102225 PubMed 40691944.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [102]Li ZQ et al. Use of Artificial Intelligence in Peer Review Among Top 100 Medical Journals. JAMA network open 2024. doi:10.1001/jamanetworkopen.2024.48609 PubMed 39625725.journal letter / editorial (full text on file; PubMed record verified) · verified 2026-09-20
- [103]Major J et al. Exploring the Endorsement and Implementation of Artificial Intelligence Guidelines in Leading Orthopaedic and Sports Medicine Journals: A Cross-Sectional Study. The Journal of bone and joint surgery. American volume 2026. doi:10.2106/jbjs.25.00373 PubMed 41706011.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [104]Shen F et al. Prevalence and Characteristics of Generative Artificial Intelligence Policies in Vascular Surgery Journals: A Cross-Sectional Review. Annals of vascular surgery 2026. doi:10.1016/j.avsg.2025.09.056 PubMed 41067523.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [105]Li W et al. Large language models for peer review in biotechnology. New biotechnology 2026. doi:10.1016/j.nbt.2026.03.007 PubMed 41903776.peer-reviewed article (full text on file; PubMed record verified) · verified 2026-09-20
- [106]Shen S, Wang K. Detecting AI-generated content in academic peer reviews. arXiv:2602.00319 (v2, 30 January 2026). arxiv.org/abs/2602.00319preprint (arXiv; full text on file; not peer reviewed) · verified 2026-09-20
- [107]Kim SSY, Deng WH, Vaughan JW, et al. Use and effects of LLMs in peer review: a randomized experiment and survey at ICML 2026. arXiv:2609.19420 (16 September 2026). arxiv.org/abs/2609.19420preprint (arXiv; full text on file; not peer reviewed) · verified 2026-09-20
- [108]Kim S, Yoon D, Gashteovski K, et al. On the limits and opportunities of AI reviewers: reviewing the reviews of Nature-family papers with 45 expert scientists. arXiv:2605.20668 (20 May 2026). arxiv.org/abs/2605.20668preprint (arXiv; full text on file; not peer reviewed) · verified 2026-09-20
- [109]Baumann J, Pei J, Koyejo S, Hovy D. Stop automating peer review without rigorous evaluation. arXiv:2605.03202 (v2, 4 May 2026). arxiv.org/abs/2605.03202preprint (arXiv; full text on file; not peer reviewed) · verified 2026-09-20
- [110]Fichtl AM, Ellinger L, Kelber J, Olík K, Groh G. AI-assisted peer review across research communities: from reviewer AI policies to LLM review quality. arXiv:2608.03581 (4 August 2026). arxiv.org/abs/2608.03581preprint (arXiv; full text on file; not peer reviewed) · verified 2026-09-20
- [111]Parsa P, Rezaei A. More than mimicking reviewers: evaluating LLMs for pre-submission peer review. arXiv:2609.05788 (5 September 2026). arxiv.org/abs/2609.05788preprint (arXiv; full text on file; not peer reviewed) · verified 2026-09-20
- [112]International Committee of Medical Journal Editors. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals: Defining the Role of Authors and Contributors; Responsibilities in the Submission and Peer-Review Process. Pages captured 20 September 2026. www.icmje.org/recommendations/browse/roles-and-responsibilities/responsibilities-in-the-submission-and-peer-peview-process.htmleditorial-body recommendations (web capture on file) · verified 2026-09-20
- [113]Committee on Publication Ethics. Case 26-02: Editors suspect that reviewers are using artificial intelligence. Advice from COPE Members, updated after the COPE Forum of March 2026. Captured 20 September 2026. publicationethics.org/guidance/case/editors-suspect-reviewers-are-using-artificial-intelligenceCOPE case advice (web capture on file; not formal COPE guidance) · verified 2026-09-20
- [114]COPE Council. COPE position: Authorship and AI tools. Last reviewed 13 February 2023. doi:10.24318/cCVRZBms. Captured 20 September 2026. doi.org/10.24318/cCVRZBmsCOPE position statement (web capture on file) · verified 2026-09-20