INTRODUCTION
The landscape of academic publishing has seen an incredible development in recent decades, influenced by the increasing production of scientific material worldwide, open access innovations, and digital platforms for spreading information. This massive growth has opened up the way for everyone to access knowledge and has brought about a lot of new channels for academic writings. However, it has also led to the creation of structural incentives and weaknesses that predatory, cloned, and fake journals have taken advantage of.1 Alongside this, artificial intelligence (AI) mainly large language models (LLMs) like ChatGPT and other generative software2 has started to change the manner in which research is imagined, drafted peer-reviewed, and shared, offering a double-edged sword: potential improvement of the research integrity but also new forms of publishing misconduct.
A. Growth of the academic publishing ecosystem
The growth of online platforms and the rise of the open access movement have contributed to a rapid increase in the number of journals, many of which are fully digital and require authors to pay article processing charges (APCs) for their publication. This approach, when combined with a thorough peer review and editorial processes, is perfectly valid but has unfortunately opened the door for the appearance of publishers and journals that focus on making money rather than producing good scholarly work, sometimes by minimizing or totally getting rid of peer review in order to maximize profits. In several countries including India, systems for governance like the UGC-CARE List have been set up mainly to inform quality journals apart from dubious3 and unethical ones and also to safeguard the academic credit system that is based on publication metrics.4
Within the enlarging ecosystem, researchers especially new or less established academics as well as those facing great publish or perish pressures will see a highly congested, less transparent and confusing environment where genuine journals do exist alongside bad or predatory and even cloned ones that imitate their looks, names, and even ISSNs. Such confusion may allow ongoing deceit and thus academic credit might be wrongly apportioned; the scholarly record may get deterioration and individual and institutional reputations may be damaged.
B. Rise of predatory and clone journals
Predatory journals are generally identified by their deceit towards authors and readers about the peer review, editorial procedures, and their presence in databases, to get papers and APCs while doing no real quality checking.5 They show fake impact factors, wrongly announce being indexed in Scopus or Web of Science or major indexing platforms, spam aggressively for papers or editorial board membership and publish basically any paper for payment, without regard for scholarly value. Predatory publishing as a term was introduced and popularized by Jeffrey Beall after he started putting out lists of such journals and publishers, thus making people around the world aware of this type of publication misconduct as a separate category.6
Clone (or hijacked) journals are a narrower and more dangerous category of unethical publications. A clone journal copies the name ISSN logo, and general appearance of a legitimate, usually well-known journal, and often runs a separate site that almost exactly replicates the real one, thus tricking authors into thinking they are submitting to the real one. Indeed, many cloned journals have the same ISSN and similar web addresses, sometimes making only the slightest changes in spelling, language, or domain name, which makes it hard for unsuspecting authors to detect them.
Measures like the UGC-CARE system have clearly named and listed cloned journals, stating that publications in such journals will not be considered for academic purposes and also that journals involved in unethical practices will be taken off the approved lists. Clone journals in an ethical sense not only involve fake misrepresentation but also stealing the identity of legitimate journals, thereby eroding the trust in scholarly communication systems.
C. Entry of AI tools in research and publishing
Generative AI tools such as large language models (LLMs), image generators, and automated summarization systems have dramatically changed how research and publishing are carried out. According to surveys, a significant number of researchers resort to AI tools at different points of their work starting from literature searches, writing drafts, to editing texts, translation, and even reference management.7 Besides, these tools can be used to boost language skills especially for those for whom English is not their first language. It can also be a helping hand when it comes to argumentation and structuring of the paper. For example, the tool can help in identifying if one plagiarized or manipulated an image. Additionally, editors and publishers are able to use these advanced tools to sift through numerous submissions in a short time, checking for various types of integrity issues.
However, AI systems present new ethical dilemmas. For instance, they may produce very believable yet fabricated information (hallucinations); AI-biases, i.e. biases in AI models coming from biases in training data, remain a big issue; new forms of plagiarism arise when the generated text is semantically derivative of existing work without the clear giving of credit. Furthermore, AI text-detection programs, although practicing, are very imperfect both in terms of false positives and false negatives, and when such tools are used as the sole basis for pinpointing a misconduct, serious due-process concerns might be raised. In response to this situation, key editorial organizations like COPE8 and the International Committee of Medical Journal Editors (ICMJE) have provided clear guidance on the ethical use of AI in writing, reviewing, and editing tasks. They have made it clear that AI is not author9 and that the extent of its usage should be openly communicated.
D. Problem Statement
This research highlights the dual role of AI in a publishing system already affected by predatory and cloned journals. While AI can be misused to generate full research papers, manipulate data and images, and bypass traditional plagiarism checks especially in low-quality journals with weak peer review it also offers strong tools for maintaining integrity. AI-based systems can detect plagiarism, analyze citation anomalies, perform image forensics, and identify fake or predatory journals by examining website features, domain history, and editorial information. Overall, AI serves both as a potential enabler of misconduct and as a powerful mechanism for detecting unethical publishing practices.
E. Review of Literature
• “False authorship: an explorative case study around an AI-generated article published under my name”10: This case study is considered a major discovery about the Global International Journal of Innovative Research (GIJIR). It revealed that among 53 studied articles, at least 48 could have been AI-generated, with a large portion being falsely credited to highly reputed authors in order to elevate the journal’s status. The report points out the extreme dangers artificial intelligence can bring to the trustworthiness of authorship. Since the study is based on only one, quite small journal, it is not very clear to what extent the results can be applied to the academic publishing sector in general.
• “Editorship Challenges in the Age of Generative Artificial Intelligence”11 This editorial section points out some real difficulties editors have had to deal with, for instance, submissions almost doubling and fake AI citations being more common. It points out that although AI technologies are great for uncovering bad practices, they are not a substitute of human decision-making, especially when the authors use AI as an accessibility aid. The paper is basically conceptually and experiential from a single journal point of view and thus lacks a thorough analysis of editor practices among different publications.
• “Hallucinated citations are polluting the scientific literature. What can be done?”12 According to this Nature analysis, the number of hallucinated citations is rapidly increasing – these are fake references produced by AI and appearing in published papers. In one review of conference papers, it was discovered that 2.6% of papers in 2025 included such references, compared to only 0.3% in 2024, which clearly shows a fast-developing issue that could undermine the reliability of scientific literature. However, the analysis is based on estimates and isolated examples without a thorough, longitudinal study that could measure the problem accurately and in detail throughout all disciplines.
• “Fabricated citations in the age of AI: A wake-up call for editors, reviewers, and authors”13 One academic editor talks about the situation where throughout a single manuscript, the majority, approximately 80%, of cited references were made up by AI however the manuscript had not only been through technical checking but also three rounds of peer review. This is a very clear warning that present methods of quality control are quite incapable to discovering submissions that are manipulated by AI. In the form of a letter to the editor, it just presents one anecdotal case and does not offer or try a methodical solution to the problem.
CONCEPTUAL FRAMEWORK
The conceptual framework of this paper is based on four key interrelated themes: (i) publication ethics as described by major normative bodies; (ii) research misconduct especially the FFP (fabrication falsification plagiarism) model and also other practices like salami slicing and duplicate publication; (iii) the differentiation of clone and predatory journals as unethical publishing models; (iv) the increasing use of AI in academic publishing both as a tool and as a risk factor. All of these notions combined serve as a theoretical basis for AI, clone journals, and publication misconduct issues and their interplay from doctrinal, empirical, and comparative perspectives.
A. Publication Ethics
1. Meaning and scope
Publication ethics is the comprehensive set of principles, standards, and protective procedures that govern the responsible communication of research findings in academic platforms. It includes not only the behavior of authors but also the editorial responsibilities reviewing publishing, and institutional duties for the assurance of the published work being legitimate clear well-referenced, and amendable in the light of new evidence or discovered mistakes in a scholarly manner and trustworthy record of the disciplines and jurisdictions. As a set of norms, publication ethics is based on a number of values such as truthfulness uprightness justice, animal and human subjects’ rights, and the sustaining of a reliable scholarly record among disciplines and jurisdictions. 14
Contemporary publication-ethics systems are often formalized by a network of regulations: journal rules, publisher codes of conduct, institutional ethics guidelines, and sector-wide frameworks issued by organizations like the Committee on Publication Ethics (COPE), the World Association of Medical Editors (WAME), the International Committee of Medical Journal Editors (ICMJE), and collaborative declarations like the “Principles of Transparency and Best Practice in Scholarly Publishing.” These systems are not characterized by a purely legal nature but operate as soft-law standards that influence expectations, provide a basis for editorial decisions (e.g. retractions), and affect the regulatory measures of funding bodies and academic authorities.15
2. Authorship norms
Authorship is at the heart of publication ethics since it is the main way through which academic accolades, obligations, and liabilities are distributed. Fundamental authorship standards dictate that individuals credited as authors must have made important contributions to the idea plan performance, or understandings of the study, be capable of publicly taking responsibility for, at least, their part of the work, and be able to recognize which co-authors are accountable for other specific parts. COPE’s core practices highlight that journals should have well-defined authorship and contributor policies, include criteria for authorship, publication acknowledgements, and dispute resolution.16
ICMJE and other similar organizations have further developed these standards when dealing with newly developed technologies, including AI tools. They have specified that AI machines cannot be authors because they cannot bear responsibility, give informed consent, or be held accountable for the integrity of the work. Concurrently, many policies require the disclosure by authors of any significant use of AI-based writing, translating, or analyzing thus reinforcing openness while maintaining human responsibility for the final paper.
3. Peer-review integrity
Peer review is considered by many as one of the most important processes in scholarly publishing. It acts as a gatekeeper by helping to determine the quality, originality, and ethical compliance of manuscripts submitted for publication. Ethical standards for publication emphasize the importance of peer review processes being well-articulated, uniformly adhered to, and properly aligned with the journal’s focus and subject area. COPE core practices promote adoption of fine-tuned peer review systems by journals, ensuring reviewer train and guidance, and upholding reviewer misconduct (e.g. plagiarism of the manuscript, submission of fraudulent reviews) handling mechanisms.17
With the advent of AI, peer review as a process is encountering the challenges of maintaining its integrity. There are accounts of the rise of AI-generated peer-review reports in which the authors of the text are the AI writers, and the reviewer only submits the text, pretending that it is their own evaluation18. This fake approach raises the issues of misrepresentation and loss of rigor, confidentiality being breached, etc. AI editorial guidelines caution using AI only as language editor or for carrying minimal checking, it is not a replacement for the expert judgement of a reviewer. Besides, any use of AI in peer review must be disclosed openly and also kept within strict confines so as to protect confidentiality and accountability).
4. Transparency and accountability
Transparency and accountability are two crosscutting principles that affect the entire publication ethics. Publishing ethically calls for journals to be transparent in publishing their aims and scope, editorial boards, peer-review processes fees conflict of interest’s policies, and indexing status, so the authors and readers are able to make informed decisions about where to submit and how to interpret the published content. Accountability on the other hand is about holding journals and publishers responsible by having corrections, expressions of concern, and retractions of the scholarly record, and also by probing allegations of misconduct in a fair and consistent manner.
The “Principles of Transparency and Best Practice in Scholarly Publishing,” co-published by COPE DOAJ OASPA, and WAME, put these ideals into action through specific conditions on journal websites governance peer review ownership archiving, and business models. In the area of clone and predatory journals, these principles are used to measure how deceptive or opaque practices are carried out and they also direct the development of algorithmic tools and checklists that help to identify suspicious journals.
5. Reference body: Committee on Publication Ethics (COPE)
COPE is a non-profit membership organisation which is devoted to giving editors and publishers guidelines, tools, and advice based on real cases for dealing with ethical issues in scholarly publishing. Its members comprise thousands of journals and publishers from all disciplines, which makes it a major hub in the worldwide governance of publication ethics. The main components of COPE are its Core Practices, flowcharts for dealing with specific types of misconduct (e.g. plagiarism duplicate publication, reviewer misconduct), discussion papers on ethical issues, and a case database that contains anonymised examples of how particular problems have been handled.
COPE’s views are very much in line with the issues around AI and clone journals. It has released position statements on AI that make it clear that AI tools cannot be listed as authors and that their use should be declared and properly limited in terms of both authorship and peer review. When it comes to doubtful journals, COPE’s transparency and core practice standards can be used as a normative reference for distinguishing legitimate journals which maintain strong policies on authorship, peer review, and ethical oversight from predatory and cloned outlets that do not meet these standards.
B. Research Misconduct
1. Fabrication, falsification, and plagiarism (FFP model)
The FFP framework, which consists of fabrication, falsification, and plagiarism, has become the leading model conceptually for describing research misconduct in many laws, and it is also widely reflected in the policies of research institutions. Fabrication is the act of inventing data, results, or citations and presenting them as though they were real; it produces false evidence that can lead other researchers and decision-makers astray. Falsification is the alteration of research components, such as materials, equipment, and procedures, or even data (e.g. by ignoring some observations, changing pictures, or reporting only positive results) to the extent that the published record does not correctly represent the evidence at the base.19
As the last part of the FFP triad, plagiarism is generally understood as the use of somebody else’s thought’s methods findings, or words without properly crediting. It can occur through verbatim plagiarism (taking the text or pictures without changes), mosaic plagiarism (patchwriting or very close paraphrasing), or conceptual plagiarism (stealing ideas or the way the work is structured), and can be related both to published or unpublished items, including grant proposals and peer-review reports. Plagiarism-related issues have recently also broadened to cover ‘self-plagiarism’ (repetitive or duplicate publication) as well as worries about AI-produced text that duplicates or closely resembles the existing literature without clear acknowledgment.
2. Salami slicing and duplicate publication
Besides FFP, the literature points out several other types of unethical publication, among which salami slicing and duplicate publication are the two most significant ones in relation to the pressures and incentives that predatory and cloned journals create. Duplicate publication occurs when almost the same data or results are published repeatedly without complete cross-reference and justification typically in different journals or languages. This practice not only misleads the scientific literature but also artificially inflates an author’s publishing productivity.
Salami slicing (or salami publication) is the term for the practice of dividing a single significant study or dataset into several minimally publishable units, each submitted as a separate article that gives only a very slight incremental contribution. While it may not always be easy to tell this apart from legitimate multiple analyses of a complex dataset, unethical salami slicing usually involves the same hypotheses, methods, and samples without enough new revelation to warrant separate papers. This practice can mislead readers about the amount of evidence, artificially raise the number of citations, and squander the time and efforts of editors and peer-reviewers.
These journals that are predatory and cloned only increase the potential for fabrications, falsifications and plagiarism (FFP), the slicing of a single scientific paper into multiple parts (“salami slicing”), and publishing the same paper more than once (“duplicate publication”). This is because they often do not have a strong peer review process, are open to receiving the same or low-quality papers, and there are instances when they even invite paper mills to produce articles for them. So, in a way, the application of AI tools in these cases may result in abuse of authors generating or repackaging content with great speed, thus further endangering the production of fragmented, plagiarized, or fabricated manuscripts.
C. Clone Journals vs Predatory Journals
1. Clone journals: fake websites mimicking legitimate journals
Clone (or hijacked) journals are fraudulent publications that imitate legitimate journals by copying their titles, ISSNs, logos, and website designs, often using slightly altered domain names to deceive authors into believing they are submitting to the genuine source. These fake journals typically target well-known, indexed publications such as those listed in UGC-CARE or major databases to exploit their reputation. Authors who fall victim may pay article processing charges and see their work appear on these illegitimate sites, which are not indexed and may disappear without notice, leaving them without valid academic credit and potentially facing scrutiny for poor judgment. Authorities like the University Grants Commission (UGC) of India regularly warn against such practices, emphasizing that publications in cloned journals will not count for academic evaluation and advising verification through trusted registries like ISSN, DOAJ, and COPE.
2. Predatory journals: low-quality, exploitative publishing
Predatory journals differ from clone journals in that they do not impersonate specific titles but instead operate under their own names while exploiting authors through deceptive practices and poor editorial standards. They often send mass email solicitations, falsely claim rigorous peer review and indexing, fabricate metrics like fake impact factors, and typically accept papers quickly with little or no review once APCs are paid, while lacking transparency about editorial boards and proper archiving or retraction systems. Although both predatory and clone journals violate ethical principles promoted by bodies like Committee on Publication Ethics, they do so differently: clone journals commit identity fraud by mimicking legitimate publications potentially involving trademark infringement and cybercrime whereas predatory journals systematically disregard ethical norms under their own branding. Increasingly, AI-based tools are being used to detect such journals by analyzing website characteristics, domain data, editorial structures, and citation patterns, helping regulators such as University Grants Commission and initiatives like UGC-CARE identify and filter out suspicious publications.
However, the issue is that predatory and cloned journals might also make use of AI in order to generate spam solicitations, counterfeit editorial content, or create AI-written articles, making detection difficult and posing new questions about liability and enforcement.
D. AI in Academic Publishing
1. AI-assisted writing
AI-assisted writing is increasingly common across academic disciplines, helping with tasks such as drafting, paraphrasing, translation, summarization, and language polishing especially benefiting non-native speakers by improving clarity, coherence, and disciplinary style. When properly disclosed and used with clear attribution of ideas and responsibility, such practices are generally accepted by editorial teams and supported by guidelines from organizations like Committee on Publication Ethics. However, ethical risks remain: AI systems can produce convincing but incorrect information (“hallucinations”), embed hidden biases, and generate paraphrased content that skirts traditional definitions of plagiarism while still relying heavily on existing work. Overreliance on these tools may also weaken authors’ critical engagement with literature and methodology, potentially reducing originality and rigor making such work more susceptible to acceptance by predatory or cloned journals that lack robust peer review standards.
2. Automated and AI-supported peer review
AI is increasingly used within editorial offices to streamline peer review through automated checks for plagiarism, image manipulation, reference accuracy, and basic methodological issues, as well as tools that help suggest reviewers, rank submissions, and provide preliminary quality assessments easing the workload on editors handling rising submission volumes. However, the use of AI by reviewers themselves raises ethical concerns: AI-generated reports may sound polished but often lack depth, and current tools struggle to reliably detect them, undermining expectations of genuine expert engagement and accountability. There are also confidentiality risks when manuscripts are shared with external AI platforms. In response, many journals guided by principles from organizations like Committee on Publication Ethics have introduced policies that either prohibit AI use in peer review or limit it to secure, in-house systems with strong data protection measures.
3. AI-generated content risks
AI can now generate complete texts, images, and datasets, raising concerns about authenticity, accountability, and the boundaries of machine-generated scholarship. This capability is increasingly exploited by paper mills and unethical authors to mass-produce articles for predatory or cloned journals, which often lack rigorous peer review, thereby inflating the volume of low-quality or fabricated research. Generative AI can also create highly realistic fake data, charts, and images that may deceive even experienced editors, especially in under-resourced journals. Although AI-detection tools attempt to identify such content, they remain unreliable, producing both false positives that risk unfairly accusing authors and false negatives that allow sophisticated fabrications to pass undetected. Consequently, ethical guidelinessupported by organizations like Committee on Publication Ethics emphasize that such tools should not be used in isolation; instead, careful human judgment, contextual evaluation, and due process are essential. In this dual role, AI acts both as a facilitator of new forms of academic misconduct and as a critical component in evolving frameworks designed to protect research integrity from threats such as predatory and cloned journals.
LEGAL AND ETHICAL FRAMEWORK
The rules that regulate AI, clone journals, and the ethics of publishing constitute a nerve-racking puzzle of intertwining international standards, national legal systems, and intellectual property laws. This piece focuses primarily on three groups of norms that, in our view, stand out for the present research: (i) international publication ethics standards, especially the ones issued by COPE, WAME, and ICMJE; (ii) India’s legal system, with special reference to the UGC-CARE programme and the 2018 plagiarism guidelines; and (iii) the legal provisions for copyright, authorship, and the legal nature of AI-created works.
A. International Standards
Committee on Publication Ethics is a voluntary body that provides widely adopted (though non-binding) guidelines for editors and publishers, including its Core Practices, Code of Conduct, and flowcharts for handling issues like plagiarism, authorship disputes, and reviewer misconduct. Its Core Practices outline key areas such as conflicts of interest, data ethics, peer review, and post-publication corrections, encouraging member journals to implement clear policies on retractions and integrity issues effectively serving as a “soft-law” standard to assess journals, including predatory and cloned ones. COPE has also begun addressing AI use, stating that AI cannot be listed as an author, must be transparently disclosed, and should be used cautiously due to risks like bias and confidentiality breaches; it also stresses transparency in journal ownership, editorial boards, and peer review as essential markers of legitimate publishing.
1. World Association of Medical Editors (WAME)
World Association of Medical Editors is a global body focused on the ethical and professional responsibilities of medical journal editors, with particular attention to clinical research, patient safety, and public health. It provides guidance on issues such as conflict-of-interest disclosure, editorial independence, trial registration, and adherence to reporting standards like CONSORT, PRISMA, and STROBE, emphasizing that editors act as gatekeepers of reliable and socially responsible medical knowledge. Regarding AI, WAME advises cautious and transparent use due to risks like biased or inaccurate outputs that could affect patient care, and it promotes alignment with frameworks from Committee on Publication Ethics and International Committee of Medical Journal Editors to support consistent global standards in authorship, peer review, and handling misconduct.
2. International Committee of Medical Journal Editors (ICMJE)
International Committee of Medical Journal Editors provides widely adopted recommendations for the conduct, reporting, editing, and publication of medical research, covering areas such as authorship criteria, conflict-of-interest disclosure, clinical trial registration, research ethics, and procedures for corrections and retractions. Its January 2024 update addresses AI use, clarifying that AI tools cannot be listed as authors since they cannot take responsibility for the integrity of the work; instead, human authors remain fully accountable and must verify AI-generated content for accuracy, completeness, bias, and plagiarism. ICMJE also warns that using AI in manuscript handling particularly by reviewers may risk breaching confidentiality if manuscripts are shared with external platforms, recommending that explicit author permission be obtained and that journals establish clear AI-use policies. These guidelines, aligned with standards from Committee on Publication Ethics, help evaluate ethical practices and highlight gaps commonly seen in predatory and cloned journals.
B. Indian Regulatory Framework
1. University Grants Commission regulations and UGC-CARE
University Grants Commission (UGC) has shifted from a metrics-driven approach to one focused on research integrity, largely in response to the rise of predatory and cloned journals. To support this, it introduced UGC-CARE, which maintained a curated list of journals eligible for academic recognition, replacing earlier “UGC Approved” lists. The framework required journals to meet standards such as genuine peer review, transparent editorial policies, verifiable indexing, and adherence to publication ethics aligning with global norms like those of Committee on Publication Ethics. Publishing outside this list generally did not count toward PhD submissions, promotions, or funding, discouraging engagement with low-quality outlets. UGC-CARE also actively identified and warned against cloned journals mimicking legitimate titles, urging researchers to verify journal details through official sources and ISSN records. Although these measures functioned more as academic regulation than legal enforcement, they played a key role in combating fraudulent publishing practices; however, the UGC-CARE List has since been discontinued.
2. UGC 2018 plagiarism regulations
The UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018 establish a formal legal framework to address plagiarism in Indian higher education, defining it broadly as presenting others’ work or ideas as one’s own and introducing a tiered system based on similarity percentages detected by software. It categorizes plagiarism into four levels: Level 0 (up to 10%) with no penalty, Level 1 (10–40%) requiring revision, Level 2 (40–60%) leading to possible rejection or loss of credit, and Level 3 (above 60%) attracting severe penalties such as cancellation of registration, with sanctions varying for students, researchers, and faculty. Although the regulations do not explicitly mention AI-generated content, their broad definition of plagiarism and reliance on similarity thresholds imply that improperly attributed AI-assisted or AI-generated material may fall under these rules, encouraging institutions to adapt policies such as mandating disclosure of AI use while reinforcing that authors remain responsible for originality and integrity.
C. Intellectual Property Laws
1. Copyright issues in academic publishing
Intellectual-property (IP) law has a few common points with publication ethics. For example, in relation to copyright and neighbouring rights of writers of articles, compilers of datasets and artists of images. Typically, the creator of a literary work has copyright in it when it is a new work and has been made permanent in a certain format. It has been customary for the author of an academic work to give away all copyright in it to the publisher when sending the work for publication, often by signing standard-form contracts which may even disallow the author from re-using or self-archiving the work.
These days open access is changing rights management and, among other things, authors are using Creative Commons licences which help them keep their copyright but give an open hand to reuse the work permanently under certain terms. Predatory and cloned journals by their behaviour misrepresent their licensing terms, publish articles without valid transfer or license agreements, or reuse contents of legitimate journals without permission facing the issue of copyright and moral rights infringement.
2. AI-generated works and ambiguity of authorship
AI-generated works create a major challenge for copyright law because traditional frameworks assume a human author. Guidance from bodies such as WIPO and national copyright offices generally follows two approaches: some jurisdictions grant copyright only to human-created works, excluding purely machine-generated content, while others recognize “computer-generated works” and assign authorship to the person who made the arrangements or inputs that led to the output. Recent discussions often distinguish three categories: fully autonomous AI outputs with minimal human input, human-directed AI-assisted creation where a human remains the main creative agent, and AI-assisted editing of primarily human-authored work. The first category most strongly challenges originality requirements, while the latter two may still qualify for human authorship if there is meaningful intellectual contribution.
In academic publishing, this creates both legal and ethical uncertainty. Where AI-generated content is not eligible for copyright, extensive reliance on AI may weaken legal protection, though ethical responsibility for accuracy and integrity still rests with human authors. Where AI is treated as a tool, normal copyright can apply, but disclosure remains important under ethical standards such as those of COPE and ICMJE, which also prohibit listing AI as an author. In contexts involving predatory or cloned journals, these issues can be exacerbated: AI-generated manuscripts may be used in paper mills to produce seemingly original but low-quality or fabricated work, while cloned journals may republish content without permission. Addressing these risks requires not only enforcement mechanisms like takedowns and legal action, but also stronger publication ethics policies and updated AI-aware regulations.
AI AS A TOOL AND THREAT IN ACADEMIC PUBLISHING
A. Positive Role of AI
1. Plagiarism detection tools and integrity analytics
AI-assisted plagiarism detection is now widely used by editors, universities, and funding agencies to support academic integrity. Tools such as Turnitin, iThenticate, and Crossref Similarity Check compare submitted manuscripts against large databases of published literature, theses, conference papers, and web content using pattern matching and increasingly machine-learning techniques to detect not only direct text overlap but also paraphrased or translated similarities. These systems help operationalize plagiarism policies, including frameworks like the UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018 and editorial standards informed by bodies such as Committee on Publication Ethics and International Committee of Medical Journal Editors, by turning abstract definitions into measurable similarity reports that support large-scale screening of submissions.
However, these tools are not definitive arbiters of misconduct and require expert interpretation, since high similarity scores may reflect properly cited material, standard methodological descriptions, or common academic phrasing, while subtle idea-level plagiarism may go undetected. Increasingly, plagiarism detection is being integrated into broader “integrity analytics” systems that also check images, verify references, and analyze citation patterns to flag potentially problematic submissions. While especially useful in high-volume environments and in detecting risks associated with paper mills or low-quality publishing practices, these systems are best understood as decision-support tools rather than replacements for human judgment.
2. Journal authenticity verification and detection of predatory or cloned outlets
AI is increasingly used not only to review manuscripts but also to evaluate the credibility of academic journals. Machine-learning models trained on datasets of legitimate and predatory journals can detect patterns such as poor website quality, lack of editorial transparency, unusual publication volumes, questionable author affiliations, and suspicious citation behaviour. These tools can assist librarians and researchers by cross-checking journal data with trusted sources, identifying false indexing claims, detecting fake editorial boards, and spotting cloned journals that mimic reputable ones. However, because errors like false positives and false negatives can occur, ethical practice requires that AI be used only as a decision-support tool, with human oversight to ensure accurate and fair judgment.
3. Research assistance and knowledge discovery
AI tools are most effective when used as research assistants that help scholars navigate and synthesize the growing body of literature rather than replace human authorship. NLP-based systems can streamline literature searches by summarizing abstracts, identifying key ideas, clustering related studies, and highlighting research gaps or emerging trends, while citation-recommendation tools suggest relevant references from large bibliographic datasets. Generative AI can also support translation, language improvement, and structural editing especially for non-native speakers promoting inclusivity when used transparently under guidelines such as those from ICMJE and COPE. Additionally, responsible AI can help researchers avoid predatory and clone journals by improving their ability to evaluate calls for papers, verify journal authenticity, and assess editorial practices, with AI-driven educational tools further strengthening research integrity awareness.
B. AI-Induced Risks
1. AI-generated fake papers and automated paper mills
The strengths that make AI a valuable research assistant also enable misuse, as large language models can quickly generate convincing but entirely fake research papers. They can produce structured abstracts, introductions, method-like sections, and discussions that mimic disciplinary norms, often including fabricated data and invented citations. Reports indicate that AI-driven “paper mills” are already using these tools to mass-produce manuscripts tailored to specific journal scopes, particularly targeting predatory and cloned journals with weak review standards. These AI-generated papers may contain “hallucinated” references either completely fake or mixed with real citations making verification difficult at scale. For journals with limited editorial resources and high submission volumes, distinguishing genuine research from AI-generated content can be extremely challenging, especially when there is little incentive to detect fraud.
2. Deepfake research data and images
AI can now generate highly realistic synthetic images, graphs, and datasets that closely resemble genuine experimental results. Techniques such as GAN-based image generation have produced deepfake western blots, microscopy images, and other scientific visuals that can even deceive expert reviewers. Investigations into paper mills suggest that some fraudulent biomedical publications may already include AI-generated images rather than traditional manipulations. Unlike older forms of image fraud, which often leave detectable traces like duplication or editing artefacts, AI-generated images can be entirely new yet plausible, making them much harder to identify. Detection tools are still evolving and often struggle to keep up with these advances, creating an ongoing challenge for research integrity. If such synthetic data are published especially in predatory or cloned journals with weak peer review they can mislead future studies, clinical decisions, and policy development. The risk becomes even greater when flawed data are included in meta-analyses or systematic reviews without proper scrutiny. In fields where replication is limited, the presence of AI-generated data could undermine the reliability of entire research areas.
C. Automation and the scaling of misconduct
AI not only increases the sophistication of research fraud but also enables it to scale rapidly. Automated systems can generate large volumes of manuscripts, spam submissions, and even fake peer reviews with minimal human effort, feeding into an ecosystem supported by cloned journals and predatory conferences that accept work with little scrutiny. Additionally, AI-powered translation and paraphrasing tools make it easier to recycle the same content across multiple publications known as salami slicing while evading plagiarism detection. In systems that prioritize publication counts, this can inflate research metrics without genuine scholarly contribution. These challenges highlight the need for broader, coordinated responses, including collaboration among publishers, regulators, funding agencies, and institutions to detect suspicious patterns, share information on fraudulent practices, and shift evaluation criteria from quantity to quality.
DETECTION AND PREVENTION MECHANISMS
A. Technological Solutions
Plagiarism detection systems have evolved from simple string-matching tools into advanced AI-driven platforms capable of understanding and comparing the meaning of texts. Tools like iThenticate and Turnitin now scan submissions against vast databases of books, articles, dissertations, and web content, producing similarity reports that highlight overlapping sections and assign percentage scores. Modern systems use machine learning to distinguish between acceptable scientific phrasing and problematic duplication, and they can also flag text that may be AI-generated based on linguistic and statistical patterns. However, these AI-detection features are probabilistic and not definitive proof of misconduct, meaning both false positives and false negatives are possible. As a result, such tools should be used as indicators for further review rather than automatic grounds for penalties. Institutions are encouraged to integrate these systems into submission workflows while also establishing clear policies on acceptable similarity levels, procedures for handling flagged content, and opportunities for authors to respond. Beyond plagiarism checks, these tools are valuable for identifying patterns linked to paper mills and detecting duplicate publications across suspicious or predatory journals, supporting broader efforts to maintain research integrity.
B. Best Practices for Researchers
1. Verifying journal authenticity: a practical protocol
For individual researchers, especially those in the early stage of their careers, having a well-organized due diligence procedure is the best measure against cloned and predatory journals. For example, Think. Check. Submit. COPE related documents, and research-based papers provide the foundation to synthesizing a detailed protocol. A thorough protocol may be as follows:
• The first step is to find out is whether one can easily find the latest issues through library catalogues, Google Scholar, or subject databases.
• Looking at the actual source of Scopus, Web of Science, PubMed (if applicable), or DOAJ instead of believing on the website logos; in India, check the UGCCARE portal.
• The publisher’s name and address should be very explicit; check that the email, phone, and physical contact information is real and is consistent with known institutions; journals whose contact details only include generic web forms or free email addresses should be treated with caution.
• Identify slight variations from the well-known titles (extra words, changed spelling, odd domains); verify ISSNs through the ISSN Portal and make sure the URL corresponds to the publisher’s official website.
• Legitimate journals offer detailed descriptions of their peer-review method including the average duration, the handling of conflicts of interest, and their affiliation with COPE or similar organizations. On the other hand, journals that guarantee acceptance, offer unrealistically short review periods, or provide vague explanations are raise suspicions.
• Find out if the editors and board members indicated in the journal are actual, currently active academics through the verification of their institutional profile pages or ORCID records. If one encounter names that are misspelled, not found, or look improbably high-level for the journal profile, be very wary.
• Find out if the APCs are visible and whether they go along with ones in similar journals. In contrast, hiding or frequently changing fees, or forcibly pushing the customer to pay immediately may be signs of predatory and cloned journals.
2. Ethical AI usage guidelines for researchers
For a single researcher, these are some practical rules that embody him:
• Generative AI should be used in very limited ways for example, when it is used for language polishing translation structuring suggestions, or preliminary idea generation the rest of the intellectual work (theoretical framing methodology interpretation) should be authentically self-developed; no credit should be given to AI systems as co-authors.
• If the policy of the journal or the institution requires the disclosure of the AI tools used and a description of the tasks to which the tools were used (e.g. grammar correction, summarizing background literature) the disclosure should be made at the acknowledgements or methods sections, as appropriate.
• Issue a detailed review of AI-generated text translations summaries, and references for exactness, bias, and originality; It’s forbidden to “blindly” trust AI-provided citations without personally verifying if the sources exist and are relevant.
• Do not employ generative models for the creation of “illustrative” data, images, or statistical that are not clearly marked as simulations or examples; If you do so, you will be technically committing the act of fabrication or falsification, even if your intention was just to be illustrative.
• Do not upload your unpublished manuscripts, sensitive datasets, or personal identifiable information to public AI platforms, unless you have explicit permission and the necessary safeguards (such as institutional licenses) are in place.
CHALLENGES
Rapid advancements in AI are making it increasingly difficult for research integrity systems to detect AI-generated or manipulated content. Although tools like iThenticate and Turnitin are widely used, they are not fully reliable and can produce false positives and false negatives, raising concerns about fairness and allowing some unethical work to go undetected. At the same time, limited institutional resources make consistent human review difficult, further weakening oversight. This challenge is intensified by the “publish or perish” academic culture, where emphasis on publication quantity and metrics encourages misconduct such as plagiarism, data fabrication, and publication in predatory journals. AI tools and paper mills have further reduced the effort needed to meet these targets, making it easier to produce or acquire questionable research in environments where ethical enforcement is weak.
CONCLUSION
Addressing the ethical challenges of AI and predatory or clone journals requires coordinated reforms across multiple levels. Legally, AI-related misconduct and journal fraud must be clearly incorporated into existing frameworks with defined responsibilities and penalties. At the institutional level, bodies such as UGC-CARE, university ethics offices, and international collaborations need to strengthen detection systems, share intelligence, and focus on managing systemic risks rather than isolated cases. Technologically, there is a need for transparent and independent tools that support human judgment, along with exploration of decentralized systems that enhance authenticity and accountability. Ethically, the global academic community must promote responsible AI use while reforming the “publish or perish” culture to prioritize quality over quantity. As AI becomes more integrated into research, misuse will continue to evolve alongside it. Instead of attempting to reverse technological progress, the focus should be on building strong socio-technical systems where AI reinforces research integrity. With aligned legal, institutional, technological, and ethical efforts drawing on global experiences it remains possible to maintain a trustworthy and open scholarly record despite ongoing advancements in AI and emerging forms of misconduct.
RECOMMENDATIONS
• Clone or fake journals pose a serious threat to research integrity, and a key solution is creating a nationally maintained whitelist of verified journals by bodies like UGC and CSIR. This list should rely on trusted indexing systems such as Scopus and Web of Science and DOI verification through CrossRef, and be used by institutions for evaluating academic work.
• Mandatory publication ethics training for students and researchers is also essential. Based on Committee on Publication Ethics (COPE) guidelines, this training should help researchers identify predatory journals, understand peer review, and avoid unethical practices, with certification linked to academic progression.
• A centralized national reporting system is needed to allow researchers to verify journals, report suspicious ones, and access updated blacklists, helping authorities take action against fraudulent publishers.
• Institutions must also be held accountable by strictly rejecting publications from clone journals in hiring, promotions, and funding decisions, with penalties for misconduct aligned with national quality assurance frameworks like NAAC.
• Finally, awareness and digital literacy should be strengthened through workshops and training on tools like DOI and ISSN verification, supported by collaborations with major publishers. Combining awareness, regulation, and technology is crucial to effectively combat clone journals.
*****
Footnotes
1. Zafer Kocak, “Publication Ethics in the Era of Artificial Intelligence” 39 Journal of Korean Medical Science e249 (2024).
2. Leila Arabi, Ali Roohbakhsh, et.al., “Artificial intelligence (AI) in academic publishing: Legitimate use, plagiarism detection, and ethical challenges” 29 Iran J Basic Med Sci 1 (2026).
3. “UGC lists 24 clone journals, teachers taken for a ride”, The Times of India, July 3, 2020, available at: https://timesofindia.indiatimes.com/city/nagpur/ugc-lists-24-clone-journals-teachers-taken-for-a-ride/articleshow/76758598.cms (last visited on April 11, 2026).
4. International Journal of Scientific Development and Research, “UGC CARE: A Battle Against Journal Clones”, available at: https://www.ijsdr.org/blog-4.php (last visited on April 11, 2026).
5. SPUBL, “Predatory and Clone Journals: How to Identify Unfair Scientific Publishing”, available at: https://spubl.com.cn/en/blog/predatory-and-clone-journals-how-to-identify-unfair-scientific-publishing (last visited on April 11, 2026).
6. F.H. Kakamad, A.M. Mustafa, et.al., “Predatory publishing lists: a systematic review” 88 Ann Med Surg 1842 (2026).
7. Thesis Edit, “AI Detection Tools, Plagiarism Checks, and Technical Writing Ethics”, available at: https://thesis-edit.com/ai-detection-tools-plagiarism-checks-and-technical-writing-ethics/ (last visited on April 12, 2026).
8. Editage Insights, “COPE Guidelines for AI Usage”, available at: https://www.editage.com/insights/cope-guidelines-for-ai-usage (last visited on April 12, 2026).
9. Neil Levy, “Responsibility is not required for authorship” 51 J Med Ethics 230 (2024).
10. Diomidis Spinellis, “False authorship: an explorative case study around an AI-generated article published under my name” 10 Res Integr Peer Rev 8 (2025). https://doi.org/10.1186/s41073-025-00165-z
11. “Editorship Challenges in the Age of Generative Artificial Intelligence” 24 J Acad Ethics 49 (2026). https://doi.org/10.1007/s10805-026-09724-5
12. Miryam Naddaf and Elizabeth Quill, “Hallucinated citations are polluting the scientific literature. What can be done?” Nature, April 1, 2026, available at: https://www.nature.com/articles/d41586-026-00969-z (last visited on April 10, 2026).
13. “Fabricated citations in the age of AI: A wake-up call for editors, reviewers, and authors” 21 J Dental Sci (2026) doi: 10.1016/j.jds.2025.09.019.
14. Pubrica, “Understanding COPE Publication Ethics”, available at: https://pubrica.com/services/publication-support/journal-selection/understanding-cope-publication-ethics/ (last visited on April 10, 2026).
15. Phcog.Net, “Core Practices COPE: Committee on Publication Ethics”, available at: https://phcog.net/archives/core-practices.html (last visited on April 10, 2026).
16. Tiffany Leung, Taiane de Azevedo Cardoso, et.al., “Best Practices for Using AI Tools as an Author, Peer Reviewer, or Editor” 25 J Med Internet Res e51584 (2023).
17. Özgür Kemal, “Artificial Intelligence in Peer Review: Ethical Risks and Practical Limits” 63 Turk Arch Otorhinolaryngol 108 (2025).
18. Dalmeet Singh Chawla, “The AI-driven research integrity crisis: How to protect the scientific record” Nature, February 20, 2025, available at: https://www.nature.com/articles/d41586-025-04032-1 (last visited on April 10, 2026).
19. “Falsification, Fabrication, and Plagiarism in Scientific Writing” Slovak Review of Borrelia and Related Infections, available at: https://www.srobf.cz/downloads/editorial/falsification_fabrication_writ_.pdf (last visited on April 10, 2026).