A leading medical ethics journal has withdrawn a recently published article after an editorial investigation identified fabricated references and evidence of irregularities in the peer review process, highlighting growing concerns about the responsible use of generative artificial intelligence (AI) in scholarly publishing.
The retracted paper, published in the Journal of Medical Ethics, examined the ethical implications of AI in the pharmaceutical industry, arguing that algorithmic bias may contribute to inequalities in healthcare delivery. The article was authored solely by Irfan Biswas, whose affiliation at the time of publication was listed as Shrewsbury Public Schools, Massachusetts, USA.
According to the journal’s official retraction notice, editors initiated an investigation after concerns were raised regarding the accuracy and existence of several references cited in the manuscript. The inquiry concluded that multiple cited sources could not be verified. The notice further stated that the author had relied on generative AI to identify and understand referenced literature but had not independently verified the cited references before submitting the manuscript.
The journal also reported finding evidence of peer review manipulation during its investigation. Although the publisher confirmed this conclusion in the retraction notice, no further details regarding the nature of the peer review concerns have been made publicly available.
BMJ Group, the publisher of the journal, confirmed that the author acknowledged being a high school student and agreed to the retraction.
The case extends beyond a single publication. Another paper authored by Biswas, published in Frontiers in Genome Editing and titled “Ethical Dimensions and Societal Implications: Ensuring the Social Responsibility of CRISPR Technology”, is currently under editorial investigation following questions concerning the author’s reported institutional affiliations. Public statements from the University of Massachusetts Chan Medical School indicated no record of the individual attending the institution, while the University of Rhode Island stated it could not determine whether a currently enrolled student with the same name was the author of the paper.
The incident reflects an increasingly recognized challenge facing scholarly publishing. As large language models (LLMs) and AI-assisted writing tools become more widely used during manuscript preparation, journals are reporting a growing number of submissions containing fabricated or non-existent references generated by AI systems. Such citation inaccuracies can undermine research reliability when authors fail to independently verify AI-generated content before submission.
The retraction also highlights the importance of robust editorial screening, careful verification of references, transparent authorship, and secure peer review systems. Publishers and editors worldwide are strengthening policies governing AI-assisted writing while emphasizing that authors remain fully responsible for the accuracy, authenticity, and integrity of all material submitted for publication, regardless of the tools used during manuscript preparation.