Paper Mills in Scholarly Publishing: A Systematic Review of Their Prevalence, Detection, and Impact on Research Integrity

Received 01 Jul, 2025

Accepted 29 Sep, 2025

Published 10 Nov, 2025

Paper mills, organized entities that produce and sell fraudulent academic manuscripts, pose a critical threat to the integrity of scholarly publishing. This systematic review synthesizes existing literature from 2010 to 2025 to examine the prevalence of paper mills, the methods used to detect them, and their overall impact on the academic landscape. Peer-reviewed studies, editorials, retraction notices, and credible reports were analyzed across databases including Scopus, Web of Science, PubMed. Results indicate a rising trend in paper mill activity, especially in biomedical and engineering disciplines. Detection techniques have evolved from manual editorial oversight to algorithmic screening tools. The consequences are profound, affecting journal credibility, institutional reputation, and public trust in science. This review underscores the need for coordinated, multi-stakeholder interventions to safeguard research integrity.

INTRODUCTION

The integrity of scholarly publishing underpins the credibility of scientific inquiry, policy-making, and academic advancement. In recent years, this integrity has come under threat from “paper mills,” which are organized entities that systematically produce and sell fraudulent academic manuscripts for profit. Unlike isolated cases of plagiarism or authorship disputes, paper mills industrialize scientific misconduct, offering fabricated data, fake peer reviews, and pre-written articles to clients seeking publication credits. These operations have introduced large volumes of low-quality or entirely falsified research into the academic record, particularly in biomedical and engineering fields^1,2.

Several high-profile retractions have drawn attention to the scale and sophistication of paper mill activities. According to Retraction Watch and investigations by major publishers, thousands of papers have been withdrawn in recent years due to suspected links with paper mills, including over 8,000 retractions from Hindawi journals alone, in addition to the closure of 19 Wiley journals following massive fraud revelations. Another study documented recurring textual and visual patterns in hundreds of papers linked to a Russia-based paper mill, and the potential of forensic provenance analysis to detect such articles³.

The Committee on Publication Ethics (COPE) and the International Association of Scientific, Technical and Medical Publishers (STM) have also issued joint guidance highlighting the structural vulnerabilities in the publication ecosystem that enable paper mills to flourish⁴.

Despite growing concern, existing literature on paper mills remains fragmented. While some studies have outlined detection techniques, including algorithmic red-flagging and image duplication software, few have attempted a comprehensive synthesis of prevalence, detection efforts, and long-term implications. Furthermore, stakeholder responses from publishers and editors to universities and funding agencies have varied in scale and rigor. A consolidated understanding of these dynamics is essential to strengthen research governance and prevent further erosion of trust in scholarly communication^5-7.

This systematic review addresses this critical gap by analyzing published studies, retraction databases, editorial reports, and policy statements from 2010 to 2025. It aims to present current knowledge about the operation, detection, and consequences of paper mills, while evaluating the measures taken by journals and institutions to counter this growing threat.

MATERIALS AND METHODS

This review follows PRISMA guidelines to ensure transparency and reproducibility.

Eligibility criteria inclusion:

	•	Peer-reviewed articles, editorials, retraction notices, and credible institutional or journal reports
	•	English language publications (or those with feasible translations)
	•	Publication between 2010-2025
	•	Explicit focus on paper mills or related detection/retraction events

Exclusion:

	•	Blog posts or opinion pieces without citations
	•	General discussions of misconduct without specific references

Search Strategy and Sources
Databases searched:

	•	Scopus
	•	Science Direct
	•	PubMed
	•	Google Scholar (for grey literature)
	•	Retraction Watch Database

Search keywords: ("paper mill" OR "manuscript factory" OR "fake papers") AND ("scholarly publishing" OR "scientific publishing" OR "academic journals") AND ("detection" OR "retraction" OR "integrity" OR "research fraud").

Data extraction and synthesis: Titles and abstracts were screened for relevance. Full-text reviews were conducted on eligible sources. A data extraction matrix was developed under key themes: Prevalence, characteristics, detection strategies, institutional responses, and impacts. Two reviewers independently coded and cross-validated the data. A PRISMA flowchart (Appendix 1) illustrates the screening process.

Appendix 1:

PRISMA flowchart

Identification:
•	Scopus (15)
•	ScienceDirect (1,115)
•	PubMed (316)
•	Google Scholar (1,850)
•	Retraction Watch (610)
•	Total = 3,906
Removed before screening:
•	Duplicates removed = 150
•	Automation tools = 0
•	Other reasons = 0
Screening:
•	Records screened = 3,756
•	Records excluded = 3,656
Eligibility:
•	Full-text reports assessed = 100
•	Reports excluded = 60 (40 not focused on paper mills, 20 insufficient data)
Included:
•	Studies included = 24
•	Reports of included studies = 24

RESULTS

Prevalence and growth: Numerous studies confirm a significant increase in paper mill activity post-2015. Retraction Watch data reveals a sharp rise in retractions linked to suspected paper mills, particularly in biomedical sciences. Countries like China, Russia, and India are frequently cited due to academic pressures, incentive structures, and systemic vulnerabilities.

Characteristics of paper mill articles: Common features include:

	•	Repetitive textual and visual patterns
	•	Authorship inconsistencies
	•	Fake affiliations or peer reviewers
	•	Implausible results or statistical anomalies

Some publishers now use machine learning tools to identify such red flags. A summary of common characteristics observed in paper mill articles is presented in Table 1.

Detection strategies: Detection methods fall into three categories:

	•	Editorial vigilance: Journal editors flag suspicious submissions based on content quality or reviewer suggestions
	•	Technological tools: Platforms like iThenticate for plagiarism, and image duplication software (e.g., ImageTwin, Proofig)
	•	Retrospective investigations: Mass retractions and audits after publication

CrossRef and COPE have urged member journals to adopt unified protocols for detection and response. Figure 1 illustrates the typical lifecycle of a paper mill-generated manuscript and the points at which detection may occur.

Institutional and publisher responses: Publishers like Springer Nature, Wiley, Elsevier, and Hindawi have retracted large volumes of suspected articles and updated their submission protocols. Universities are reassessing promotion criteria and emphasizing research quality. Funders are implementing stricter compliance checks and audits.

Fig. 1: Paper mill lifecycle and detection points

Table 1:

Summary of key characteristics of paper mill articles

Characteristic	Description
Textual patterns	Repetitive phrasing, generic templates, and poor transitions
Visual indicators	Duplicated or manipulated images; generic or recycled figures
Authorship inconsistencies	Fake affiliations, ghost authorship, unverifiable co-authors
Statistical anomalies	Implausible p-values or perfect results; lack of raw data
Citation behavior	Excessive self-citations; citation of obscure or unrelated literature
Reviewer fraud	Faked reviewer identities or email addresses; fraudulent peer review process

Impact on research integrity: The consequences are severe:

	•	Undermining of journal reputation and public trust
	•	Citation contamination in legitimate studies
	•	Resource waste and research misdirection
	•	Reputational harm to early-career researchers

In some cases, fabricated findings have been cited in legitimate policy or clinical contexts, amplifying the risks.

DISCUSSION

This review provides compelling evidence that the expansion of paper mills represents a systemic and deeply embedded threat to the reliability of academic publishing. The major findings highlight that paper mill activities are no longer limited to fringe journals or isolated cases but have infiltrated mainstream scientific publishing through increasingly sophisticated methods. We observed recurring patterns in retracted articles, such as image duplication or manipulation, recycled textual content, and fictitious authorship, which suggest intentional fabrication supported by commercial services. More critically, the review underscores that despite improvements in detection technologies, many paper mill-generated articles continue to circulate and influence subsequent research undetected.

Current analysis reveals that detection efforts, while growing, remain unevenly distributed across publishers and disciplines. Journals with higher editorial capacity have begun implementing artificial intelligence tools to screen for textual and visual anomalies, yet smaller or rapidly expanding journals often lack such infrastructure. Moreover, editorial interventions typically occur after publication, meaning many fabricated articles are retracted only years later, by which time they may have already been cited and incorporated into academic discourse. About 44.8% of paper mill articles were still being mentioned in Q2-ranked systematic reviews, and provenance tracking was used to identify internal patterns across fraudulent submissions. Therefore, traditional peer review processes remain vulnerable to manipulation when unsupported by forensic or technological tools because they rely heavily on trust and human judgment, which can be compromised⁸.

One of the most significant vulnerabilities is the potential for reviewer fraud. This can occur when authors suggest fake reviewers or others with whom they have a relationship. These “peer reviewers” can then provide a positive, uncritical review, fast-tracking the manuscript’s acceptance regardless of its quality or scientific merit. Another vulnerability is the manipulation of citations. Authors may be pressured to include citations to papers from a specific journal or even to the work of the editor or reviewers, in a practice known as “citation cartels”. This inflates the citation counts of certain individuals or journals, creating a false impression of their influence and impact⁹.

Another major issue is the lack of transparency and accountability in traditional peer review systems. In these models, the identities of reviewers and authors are hidden from one another, which can be beneficial but also creates an environment where misconduct is difficult to detect. Without a robust system to verify the identity and legitimacy of reviewers, it becomes easy for fraudulent actors to exploit the process. Furthermore, the manual nature of the process, which often relies on email correspondence and human discretion, lacks the automated checks and balances that could flag suspicious behavior^10,11.

Paper mills operate through recurrent templates and fabricated reviewer identities, enabling the fast production of seemingly credible submissions¹². Current findings also align with the mass retractions that the closure of entire journals and withdrawal of thousands of articles, particularly within the Hindawi-Wiley portfolio. These mass retractions point to a reactive rather than proactive system, where exposure depends on whistleblower activity or post-publication audits¹³.

Recent studies have demonstrated that reliance solely on traditional peer review is insufficient to identify fake or paper mill-generated manuscripts. A detailed analysis of metadata from twelve retracted fake papers found that these manuscripts closely resembled legitimate submissions in multiple aspects, such as submission timelines, citation metrics, and author affiliations, making pre-publication detection particularly challenging¹⁴.

Only a few subtle cues, such as the consistent use of non-institutional email addresses and the excessive citation of obscure or low-quality literature, were found to be somewhat indicative of fraudulent origins. These limitations highlight the urgent need for proactive and preemptive screening strategies, rather than relying on post-publication detection and retraction¹⁵.

In response to these detection challenges, an innovative approach that focuses on the analysis of co-authorship networks to uncover fabricated publication patterns. Their method revealed structural anomalies, including unusually low clustering coefficients, star-like hub formations, and artificially constructed authorship histories¹⁶.

Such anomalies suggest that paper mills manipulate not only textual and visual content but also fabricate entire scholarly social networks to create the appearance of legitimate academic collaboration. This level of deception increases the difficulty of detection and raises the stakes for maintaining research integrity.

To counter these tactics, integrating multiple complementary methods is recommended. Metadata analysis⁷ can identify unusual submission or authorship patterns; authorship network forensics can reveal collaboration anomalies; and linguistic anomaly detection can highlight repetitive or formulaic writing styles typical of paper mill outputs.

By combining these tools into a multi-layered screening strategy, the scholarly community could significantly improve the early identification of fraudulent manuscripts, reducing the reliance on retractions as the primary corrective measure. Such proactive measures would help protect the integrity of the scientific record, maintain public trust in academic publishing, and safeguard the reputations of legitimate researchers¹².

Nonetheless, this review has certain limitations. The synthesis was constrained to English-language publications, potentially overlooking regional documentation or investigations published in other languages. Additionally, because the problem of paper mills is still emerging, much of the current literature is descriptive, with limited empirical or intervention-based studies. Our reliance on publicly reported data may underrepresent the true scale of the problem, as many journals do not transparently disclose reasons for retraction or editorial rejection. There is also limited evidence on the effectiveness of recently proposed detection tools, as many are proprietary or not peer-reviewed¹⁷.

Technological and forensic tools have emerged as crucial countermeasures against these vulnerabilities. Tools for plagiarism detection (e.g., Turnitin, iThenticate) are now standard practice, flagging text that has been copied from other sources. Data and image manipulation detection software can analyze figures and data for signs of alteration, duplication, or fabrication. These tools help to ensure the integrity of the research findings themselves. Also, advanced reviewer identification and verification systems are being developed. These platforms use algorithms and databases to check for conflicts of interest and to verify that suggested reviewers are legitimate experts in the field. Some platforms even use artificial intelligence to match manuscripts with the most suitable, unbiased reviewers. The use of blockchain technology has also been proposed to create an immutable, transparent record of the peer review process, making it nearly impossible to alter review histories or forge reviewer identities⁹.

Going forward, there is an urgent need for cross-publisher data sharing on suspected paper mill submissions, implementation of open-access forensic screening tools, and reform of academic reward structures that contribute to fraudulent publication demand. Institutional policies should emphasize transparency in authorship contributions, limit excessive authorship inflation, and provide formal protections for whistleblowers reporting suspected misconduct. Without coordinated, system-wide reforms, the scholarly community will remain vulnerable to organized deception, ultimately eroding the public trust in science that publishing exists to uphold.

CONCLUSION

Paper mills represent a systemic vulnerability in academic publishing. Their proliferation has far-reaching implications for research integrity and credibility. While progress has been made in detection and response, a concerted, global strategy is essential. Only through collaboration among journals, researchers, institutions, and funders can the integrity of scientific literature be preserved.

SIGNIFICANCE STATEMENT

This systematic review highlights the growing threat of paper mills to scholarly publishing and research integrity. By synthesizing recent evidence on their prevalence, detection, and impact, the study provides essential insights for editors, institutions, and policymakers seeking to safeguard scientific credibility and improve fraud prevention mechanisms in academic publishing.

REFERENCES

1. Cardenuto, J.P., D. Moreira and A. Rocha, 2024. Unveiling scientific articles from paper mills with provenance analysis. PLoS ONE, 19.
2. Tang, G. and H. Cai, 2025. Citation contamination by paper mill articles in systematic reviews of the life sciences. JAMA Netw Open, 8.
3. Wittau, J. and R. Seifert, 2024. How to fight fake papers: A review on important information sources and steps towards solution of the problem. Naunyn-Schmiedeberg's Arch. Pharmacol., 397: 9281-9294.
4. Parker, L., S. Boughton, L. Bero and J.A. Byrne, 2024. Paper mill challenges: Past, present, and future. J. Clin. Epidemiol., 176.
5. Wittau, J. and R. Seifert, 2024. Metadata analysis of retracted fake papers in Naunyn-Schmiedeberg’s Archives of Pharmacology. Naunyn-Schmiedeberg's Arch. Pharmacol., 397: 3995-4011.
6. Cogan, E., 2022. Preventing fraud in biomedical research. Front. Cardiovasc. Med., 9.
7. Porter, S.J. and L.D. McIntosh, 2024. Identifying fabricated networks within authorship-for-sale enterprises. Sci. Rep., 14.
8. Candal-Pedreira, C., J.S. Ross, A. Ruano-Ravina, D.S. Egilman, E. Fernández and M. Pérez-Ríos, 2022. Retracted papers originating from paper mills: Cross sectional study. BMJ, 379.
9. Cooke, S.J., N. Young, K.S. Peiman, D.G. Roche and J.C. Clements et al., 2024. A harm reduction approach to improving peer review by acknowledging its imperfections. FACETS, 9.
10. Proctor, D.M., N. Dada, A. Serquiña and J.L.E. Willett, 2023. Problems with peer review shine a light on gaps in scientific training. mBio, 14.
11. Lisciandra, C., 2025. Citation metrics: A philosophy of science perspective. Episteme.
12. Horbach, S.P.J.M. and W. Halffman, 2019. The ability of different peer review procedures to flag problematic publications. Scientometrics, 118: 339-373.
13. Fanelli, D., 2013. Why growing retractions are (mostly) a good sign. PLoS Med., 10.
14. Bik, E.M., A. Casadevall and F.C. Fang, 2016. The prevalence of inappropriate image duplication in biomedical research publications. mBio, 7.
15. Ferguson, C., A. Marcus and I. Oransky, 2014. Publishing: The peer-review scam. Nature, 515: 480-482.
16. Moher, D., L. Bouter, S. Kleinert, P. Glasziou and M.H. Sham et al., 2020. The Hong Kong principles for assessing researchers: Fostering research integrity. PLoS Biol., 18.
17. Resnik, D.B., T. Neal, A. Raymond and G.E. Kissling, 2015. Research misconduct definitions adopted by U.S. research institutions. Account. Res., 22: 14-21.