The editor received a letter from “Dr. BS’ Many other editors did the same.

Letters to the editor from authors using chatbots are flooding academic journals around the world, new research and journal editors show.

This practice threatens a part of scholarly publishing that editors believe is necessary to sharpen research findings and create new research directions.

A new study into this problem began with a tropical disease specialist who had a strange experience with a letter written via chatbot. He decided to find out exactly what was going on and who was submitting all these letters.

The scientist Dr. Carlos Chaccour, of the Institute for Culture and Society at the University of Navarra in Spain, said his investigations began immediately after the publication of an article in the New England Journal of Medicine, one of the world’s most prestigious journals. The paper, published in July, looked at controlling malaria infections with ivermectin and featured a laudatory editorial.

Then, 48 hours later, the magazine received a strongly worded letter. The editors considered publication and, as usual, sent it to Dr. Chaccour.

“We wish to raise strong objections,” the letter began, going on to say that Dr. Chaccour and his colleagues did not refer to a groundbreaking 2017 paper that showed mosquitoes become resistant to ivermectin.

Dr. Indeed, Chaccour was well aware of the “groundbreaking work.” He and a colleague had written it, and it didn’t say that mosquitoes became resistant.

The letter went on to say that an economic model showed that the method would not work to combat malaria.

Here too the reference referred to an article by Dr. Chaccour and colleagues.

“Me again? Really?” thought Dr. Chaccour. This paper did not say that the method would not work.

“It has to be AI,” decided Dr. Chaccour.

An extensive language model must have been used to write the letter, argued Dr. Chaccour. He believes that while looking for references in a niche area where there aren’t many, he looked into two of Dr. Chaccour’s own work has surfaced.

He told the diary what he found. The letter was not published.

Dr. Eric Rubin, the magazine’s editor-in-chief, said he hadn’t thought about the possibility of chatbots writing letters until he was told by Dr. Chaccour’s experience heard.

There’s a reason authors might turn to AI, noted Dr. Ruby in an interview.

Letters to the editor published in academic journals are listed in databases that also list journal articles, and Dr. Rubin said that “they count as much as an article.”

“For a very small amount of work, someone can add an article in the New England Journal of Medicine to their resume,” he said.

“The incentive to cheat is high,” he added.

Dr. Chaccour had to ask himself: Who was this person who sent the letter?

He found that the author was a doctor from a Middle Eastern country who had not published letters to editors of scientific journals until 2024.

Suddenly, in 2025, he published 84 letters on 58 topics.

“He’s a Leonardo,” said Dr. Chaccour.

Dr. Chaccour wanted to write about his experiences. He wanted to use the doctor’s initials to identify him, but realized that wouldn’t work: his initials are BS. He didn’t think he had a Dr. to be able to write BS, so he named him author #1 in a report that examined the proliferation of letters to the editor after the end of 2022, when AI became widely available.

The study included an analysis of more than 730,000 letters to magazine editors published since 2005. They were published online before being submitted to a peer-reviewed journal.

“Something happened in 2023,” said Dr. Chaccour. There was a sudden emergence of authors who had previously published few or no letters, but who suddenly had letters appearing regularly – a path that, as Dr. Chaccour said he went “from zero to heroes.”

An author from a Southeast Asian country published 234 letters in 2024 and 243 on October 31 this year, after publishing none in 2023.

He also identified 128 authors who had never published a single letter. Then, in their first year of letter writing, they had at least ten published letters.

Up to 3,000 authors who had never published a letter before 2023 have published at least three, he said.

“When someone comes out of the blue and writes three, five or 10 letters in a year, it causes a stir,” said Dr. Chaccour. “To write a letter you need expertise – you really have to be up to date with the literature.”

Dr. Amy Gelfand, editor-in-chief of Headache magazine, has started receiving suspicious letters. One clue, she said, is letters that arrive a few days after an article is published. Human writers, she said, usually take a few weeks.

She started searching for the authors of questionable letters in PubMed, a database of academic publications. An author of a recent letter published six letters to the editor in six magazines on six topics this month.

Keith Humphreys, deputy editor of Addiction magazine, received a seemingly sensible letter to the editor. He sent it to the paper’s authors requesting comments.

It turned out that the Chinese-based authors were extremely prolific. Within six months they had published letters to the editors of journals in cardiology, emergency medicine, endocrinology, gastroenterology, hepatology, immunology and critical care.

“They mastered every single area,” said Dr. Humphreys.

The number of suspicious letters continues to increase, Dr. Chaccour and his colleagues. In 2023, the proportion of letters from prolific authors – those who had published three or more in a year – was 6 percent. In 2024 it was 12 percent. This year, investigators report, it is approaching the 22 percent mark.

They are invading magazines “like Omicron,” said Dr. Chaccour, referring to the Covid variant that quickly became dominant.

The situation “is not good,” said Dr. Ruby. But the answer is not to stop publishing letters.

“Sometimes letters contain important information,” he said. “Good letters ask good questions or raise points that the writers don’t address.”

Without letters, said Dr. Gelfand, “you are missing all the value, the new findings and important criticisms and discussions about what they mean for science.”

Another idea is to stop indexing the letters so they don’t appear in PubMed.

That’s not a good solution either, said Dr. Gelfand. Links in PubMed from letters are helpful for those doing research.

There is currently no agreement on what to do.

Dr. Chaccour said that his experience with the letter from Dr. Although BS is funny, the overall picture is not.

“It’s frightening,” he said.