Artificial Intelligence (AI) Generated "Letters to the Editor" - Helpful Tool or Hidden Hazard? Experience of the PRiMER Journal
A recent letter to the editor highlighted that, with artificial intelligence (AI) able to summarise text and draft manuscripts, people can now generate multiple letters using AI to critique papers and generate responses. This phenomenon has concerned journal editors.
This letter to the editor was published in November 2025, in the Journal PRiMER
Editorial Experience With an AI-Generated Letter to the Editor
The PRiMER editorial team encountered such a case in 2025 and believes that it is important to report this emerging issue. A submitted letter commenting on a recently published paper caught the attention of the editor-in-chief for being unusually critical. While letters typically focus on specific issues, this submission appeared to be a critical appraisal exercise that repeated the limitations already noted in the original publication. Despite the authors' attestation that they did not use AI tools, three AI-detection tools indicated that 46%-100% of the content was AI-generated. The investigation revealed that the authors' ORCID profiles contained nearly 100 critical letters in various fields.
Concerns Observed by the PRiMER Editorial Team
While AI-generated letters may identify flaws in a particular paper, a careful human critique that highlights undisclosed limitations adds valuable insights to the literature. An unchecked AI-generated critique that ignores context risks being useless or misleading, potentially undermining readers' trust in the original publication. Letters that slip past editorial desks with less scrutiny than research articles crowd the scientific literature. Allowing AI-generated, poorly monitored letters increases noise and erodes the signal-to-noise ratio in the literature. The rapid generation of meaningless or erroneous letters, which endure in the scholarly literature, degrades publication value. While many debate the merits of assessing scholars using various publication metrics, peer-reviewed publications remain a de facto part of academic culture. The proliferation of AI-generated letters may render the peer review and publication processes less meaningful for end readers.
Learning for Medical Researchers, Reviewers & Related Stakeholders
Given that letters to the editor are part of the indexed biomedical literature and are frequently consulted by real-world clinicians for decision-making, the unchecked proliferation of AI-generated critical letters, produced without substantial human scholarly expertise, may complicate the interpretation of research findings and impose additional responsibilities on editors to maintain the integrity of clinical discourse. This also poses challenges for journal editors, academic disciplines, and real-world clinical practice. Fellow editors and scholarly publications may like to remain vigilant regarding this novel and challenging application of AI.
Reference: Morley CP, Kost A, Barreto T, et al. An Emerging Issue in Journal Publishing: The Use of Artificial Intelligence to Generate Letters to the Editor. PRiMER. 2025;9:65. Published 2025 Nov 14. doi:10.22454/PRiMER.2025.391630
