129 papers retracted, and recurrent use of a single sentence in innumerable papers by Indian researchers
In January-February this year, Neurosurgical Review, a Springer Nature journal, retracted 129 Commentaries. Of them, 87 Commentaries are by the Chennai-based Saveetha Institute of Medical and Technical Sciences (SIMATS). Among the 87 retractions from Saveetha University, 35 Commentaries are jointly written by Dr. Hethesh Chellapandian and Dr. Sivakamavalli Jeyachandran, while Dr. Chinnasamy Ragavendran is an author/coauthor in 21 Commentaries. There are a handful of researchers from Saveetha with 7-10 retractions each. On February 10, 2025, Retraction Watch was the first to report this.
Two other Indian-based institutions in Wardha, Maharashtra — Datta Meghe Institute of Higher Education & Research, and Mahatma Gandhi Institute of Medical Sciences — too have a large number of retractions from the same journal. There is at least one author from Datta Meghe Institute and Mahatma Gandhi Institute in 37 and 33 retracted Commentaries, respectively. Twenty-three retracted Commentaries are coauthored by Datta Meghe Institute and Mahatma Gandhi Institute researchers.
Among foreign universities, Université Technologique Bel Campus, University in Kinshasa, in the Democratic Republic of Congo has the highest at 16, and Dr. Tangmi Djabo Eric Adrien in an author/corresponding author in all the 16 Commentaries. Researchers from Mahatma Gandhi Institute and Datta Meghe Institute are coauthors in 16 and 12 of these Commentaries, respectively.
Of the 37 retractions from the Datta Meghe Institute, 36 are by Dr. Mayur Wanjari. He is a corresponding/co-corresponding author in almost all the 36 Commentaries. Dr. Roshan Prasad from the Datta Meghe Institute is an author/corresponding author in 29 of the 37 retractions, while Dr. Gaurav Mittal from the Mahatma Gandhi Institute of Medical Sciences is an author/corresponding author in 31 of the 33 retracted Commentaries.
In December 2024, an investigation by the journal Science and Retraction Watch found that a couple of journals, including the journal Neurosurgical Review, were “overwhelmed” by Commentary articles. The joint investigation found that “letters, comments, and editorials comprised 58% of the total output from January to October” 2024. And over “80% of these commentaries are from South Asian countries, compared with fewer than 20% of research and review articles”.
Reasons for retraction
The retraction notices on all these papers read the same: “The Editor-in-Chief has retracted this article. An investigation by the Publisher found that a number of articles, including this one, that comment on research published in the Journal were submitted over a short space of time and show strong indications that the text was generated by a large language model (LLM) without proper disclosure by the authors. These articles are therefore in breach of the Journal’s editorial policy and are being retracted.” As the retraction notice says, the Editor-in-Chief and the publisher had already decided to retract the Commentaries and all the authors were only informed of the journal’s decision and the content of the retraction notice.
Admitting to using an “AI tool (ChatGPT) mostly for editing and refining the language of the published article”, Dr. Jeyachandran from Saveetha University says in an email to the journal: “I realised that I did not disclose the use of an AI language model, specifically ChatGPT, in the preparation of the manuscript. This oversight was due to a lack of awareness at the time of submission, and I now understand the importance of transparency in this regard.” Apologising for the oversight, Dr. Jeyachandran had requested that an erratum be published by citing similar instances in journals by other publishers, which was not entertained by the journal.
“We have used AI-based tools to enhance the language clarity and structural presentation of the manuscripts (as per the journal policy). These tools were employed purely for linguistic refinement, not for generating scientific ideas. I’ve not done anything against the research ethics (i.e., I’ve not copied any research data or part of literature from published articles or faked my data),” Dr. Jeyachandran says in an email to me. “We gave a clear explanation, but the integrity team was not willing to accept our explanation; rather, they replied that they had already decided on the retractions.”
Claiming that all Commentaries were “accepted outright without any revisions or questions”, the decision to retract them “without first seeking our explanation or providing us with an opportunity to respond is both unexpected and, in our view, inappropriate”, Dr. Jeyachandran says. However, the author submission guidelines of Neurosurgical Review clearly mention that the use of an LLM should be disclosed in the text.
Wardha researchers deny using LLM
Refuting that the Commentaries were generated using an LLM, Dr. Wanjari from the Datta Meghe Institute says in an email to me: “We approached the journal multiple times in response to their concerns, but we have not received a single reply to any of our emails to date.” He goes further to say: “We submitted official Turnitin reports (paid version) showing 0% AI-generated content, confirming that no LLM or AI tools were used in the preparation of the manuscript. Despite this, the journal proceeded with the retraction without addressing our clarifications — and notably, without even providing any AI-detection report to substantiate their claims. We were only informed of the retraction via email, without any detailed justification or evidence.” Samples of Turnitin reports certifying 0% AI content in two retracted papers by the Wardha-based institutions were shared with me.
“According to the journal’s own guidelines, tools like Grammarly and Paperpal fall under acceptable AI-assisted editing platforms that do not require disclosure. These were the only tools employed during manuscript preparation,” says Dr. Mittal from Mahatma Gandhi Institute in an email. The journal’s submission guidelines on LLM use say: “The use of an LLM (or other AI-tool) for “AI assisted copy editing” purposes does not need to be declared. In this context, we define the term “AI assisted copy editing” as AI-assisted improvements to human-generated texts for readability and style, and to ensure that the texts are free of errors in grammar, spelling, punctuation and tone.”
“We submitted a clear denial of LLM usage, offered to provide signed declarations from all co-authors, and expressed readiness to undergo independent verification,” says Dr. Prasad from Datta Meghe Institute in an email. “The retractions cite “strong indications” of LLM use without sharing any specific evidence, and the claim that authors did not respond is demonstrably false.”
Documents show that both the Wardha-based institutions had responded to the journal and disagreed with the retractions in all the cases. Yet, 23 retraction notices say that the “authors have not responded to correspondence regarding this retraction”.
Submission times
The retraction notice says the submission of the Commentaries within a short time frame strongly indicates that an LLM was used to generate the text and the same was not disclosed by authors. All the 34 Commentaries by Dr. Chellapandian and Dr. Jeyachandran from Saveetha University were published in just a month between September 14, 2024 and October 14, 2024. Dr. Jeyachandran contests this. “The integrity team claims that we have published within a short period, which is not true… [They were] submitted to the journal at different points in time. When we submitted the Commentaries, all were accepted at a single stretch, about which we were more surprised,” he says.
Similarly, the majority of submissions by the two Wardha-based institutions were during September last year, with only one in July, five in August and eight in October 2024. “The submissions were part of a coordinated academic initiative involving multiple collaborators across institutions. These brief, literature-based commentaries were developed over time and submitted during a window when our academic schedules aligned, and journal responsiveness was noted to be higher. Their concise format made simultaneous preparation feasible. Each manuscript was individually reviewed and submitted in accordance with the journal’s policies, and we were not informed of any restriction on submission volume or frequency,” says Dr. Wanjari from the Datta Meghe Institute.
Besides being the Research Director at the Association of Future African Neurosurgeons (AFAN) since 2024, Dr. Adrien from the Université Technologique Bel Campus, DRC is also in charge of coordinating the Research Fellowship that usually starts in February each year. “In 2024, during the fellowship, the fellows were asked to write a commentary on the state of neurosurgery in their respective countries. If you have seven fellows in training during a fellowship and each of whom is given the task of writing a commentary on the state of neurosurgery in their country, we will end up with seven commentaries to submit at once. This is what happened with five articles,” he says in an email. “The journal didn’t seem to be writing to me to understand anything. The journal’s decision had already been made. So I was informed of this decision to retract our articles published several months earlier.”
#####################
In June 2023, the journal Science carried an article based on an investigation by Retraction Watch that found that Saveetha University had resorted to large-scale citing of unrelated papers published by other Saveetha University groups, which is called institutional self-citation, to “inflate the number of citations”.
In 2022, a study looking at self-citation by a few universities in Tamil Nadu during the period 2017-2021 found Tamil Nadu Physical Education and Sports University had 100% self-citation, followed by 40% for Tamil Nadu Open University, 11.71% for Alagappa University, 8.85% for Anna University, and 7.83% for Annamalai University. The study was published as a chapter in a book (Recent perspectives on Technology transfer in Library Science Education). According to a preprint (which is yet to be peer-reviewed) posted on May 12, 2025 in BioRxiv, at 16%, self-citation by Chitkara University in Chandigarh is highest for the top 2% most cited articles. It is followed by Saveetha University with 12%, King Saud University in Saudi Arabia with 9% and King Khalid University in Saudi Arabia with 8%. In comparison, self-citation by IISc is just 3%.
A single sentence
A single 15-word sentence is reproduced verbatim, including a typo/grammatical mistake, in at least a couple of hundred papers published by Saveetha University faculty. The sentence reads: “Our team has extensive knowledge and research experience that has translate [sic] into high quality publications.” The sentence is followed by citations to publications by other Saveetha groups that very often appear unrelated to the published paper. Many more papers carry the same sentence but with the typo/grammatical mistake corrected.
Several other papers carry variants of the single sentence: “Our research and knowledge have resulted in high-quality publications from our team” and “Previously our department has published extensive research on various aspects of prosthetic dentistry.” Some papers carry the corrected sentence and the grammatically wrong sentence one following the other. In all these cases, the sentence is followed by citations to other Saveetha publications.
While a research group may routinely use certain words, phrases or even a sentence or two in the methods section and to acknowledge the support provided by the institution and/or the funding agencies, a single sentence appears in innumerable papers published by different groups from the University almost always in the main text. While citing papers published by the same group will still amount to self-citation unless and otherwise the published paper builds on the research carried out in the past and those cited papers are relevant to the current paper, citing papers from other groups, and that too papers that appear unrelated, is contrary to the claim made in the sentence: “Our team has extensive knowledge and research experience that has translate [sic] into high quality publications.”
Extreme examples of citing unrelated Saveetha publications include a paper that compares the efficacy of commercially available chemical and herbal mouthwash against oral microorganisms published in December 2021 by researchers from Saveetha Dental College and the Department of Microbiology at the Saveetha Dental College and Hospitals citing a June 2016 paper by a Saveetha University coauthor from the Department of Mechanical Engineering on the performance, emission and combustion characteristics of single cylinder diesel engine powered by Cymbopogon flexuosus biofuel published in the journal Energy Conversion and Managment. Another extreme example is a December 2021 paper on the effect of Ocimum sanctum biocompounds against a gene protein of Acinetobacter baumanii citing the same 2016 paper on a single cylinder diesel engine. The 2016 paper is cited by yet another December 2021 paper studying the effects of sleep deprivation in students.
“The reproduction of a sentence, with a characteristic misspelling, across so many papers, cannot be accidental. The fact that this sentence is invariably followed by a large number of citations to other papers from the university, often with no direct connection to the paper, is suggestive of a systematic attempt to self-cite. This violates what we conventionally think of as ethical behaviour,” Dr. Gautam Menon, Professor of Physics and Biology at Ashoka University says in an email.
Origin explained
Explaining the origin and recurrent use of the sentence with a typo or its variants in papers published by Saveetha faculty, Dr. Sheeja S Varghese, Professor and Registrar of SIMATS in an email, says: “The sentence in question, including the citations that follow, was originally used by a specific research group at Saveetha that was guiding a large number of scholars as part of an extensive collaborative network. Their publications reflected this shared authorship and mentorship model. As a result, some overlap in language appeared across papers, especially in introductory sections, which may explain the recurrence of this sentence or its variants. However, this was never intended to mislead or to artificially inflate citations from [the] institution.”
Dr. Varghese further adds: “Following the earlier reporting of this issue by Retraction Watch three years ago [June 2023], Saveetha University conducted a comprehensive internal investigation and [sought an] explanation from researchers. Our findings confirmed that the majority of citations were in fact relevant and scientifically appropriate. Nonetheless, we did identify some inconsistencies in a small number of manuscripts. These were promptly addressed. The authors involved were instructed to take corrective measures wherever required, and since that time — over three years ago — we have not encountered similar instances.”
The sentence in 2025 papers
Justifying the citations of other Saveetha publications in her paper, a corresponding author from Saveetha University also says in an email that no articles from Saveetha with such sentences have appeared in the last two years. However, contrary to the claims made, Google Scholar throws up several papers published this year and in 2024 carrying the problematic sentence or its variations. In 2025, the journal Cuestiones de Fisioterapia published nearly two-dozen papers mentioning the sentence or its variants and self-citations immediately following it. While a paper published this year in the same journal has used the corrected sentence, there are no citations to Saveetha publications. Also, there are also several papers published this year in the same journal by Saveetha without the problematic sentence or its variants and without self-citations, thus confirming that corrective measures were indeed undertaken by the University to address the problem.
Explaining the reasons why about two-dozen papers published this year still carry the problematic sentence, Dr. Varghese says in an email: “While some of the mentioned articles appear to bear 2025 publication dates, these manuscripts were in fact written and submitted in 2023.” In response to a question on why the corresponding authors failed to remove the problematic sentence and the self-citations from the article with an explanation after the papers were submitted in 2023 and before they were published in 2025, the Registrar says: “We were not aware [of papers published in 2025 still carrying the sentence and self-citations to Saveetha publications]. We will investigate. However, it is important to note that these papers were authored in 2023, and many of the authors are no longer affiliated with Saveetha.”
Clarifying the issue on self-citations, Dr. Varghese says: “Saveetha maintains the highest research and publication standards in India. We are proud to report that over 90% of our published articles appear in Q1 and Q2 journals, making us the top-ranking institution in the country by quality — nearly twice the output of the next leading university. Our citation index is also among the highest, even when institutional self-citations are fully excluded, as they are by all major ranking agencies. Hence, the institution derives no benefit from, nor does it endorse, widespread self-citation practices.”
She adds: “At the time of the incident [2023], our overall self-citation rate in the dental research domain was approximately 16%, which compares favourably with other leading global institutions such as Harvard, which had a self-citation rate of 21%. Moreover, institutional self-citations are excluded from most international ranking methodologies and thus confer no advantage in terms of academic reputation or ranking.”
