JAMA® and the JAMA Network of 13 peer-reviewed journals have adopted a new policy regarding authors’ use and disclosure of augmented intelligence (AI) tools aimed at addressing a sharp rise in “AI slop” submissions from authors, particularly those submitted as letters to the editor or as opinion essays.
“We're starting to see what people colloquially refer to as AI slop. And we see that AI slop gumming up our systems,” said Kirsten Bibbins-Domingo, MD, PhD, MAS, the editor-in-chief of JAMA and the JAMA Network. “The reason behind the updated guidance is not that we're no longer bullish on AI, but we just want authors to be clear on how they should use it.”
Early in 2023, in the wake of the first burst of excitement about generative chatbots such as ChatGPT, JAMA’s editors made clear their stance that nonhumans could not be listed as authors of articles published in the journal.
That bedrock stance continues—and was furthered with detailed directions for authors in 2024—but the newly updated JAMA guidance sets out clearer rules of the road on how authors can and cannot use AI tools, as well as how they must disclose and describe that use.
Six AI no-no’s for JAMA authors
“We as editors embrace AI as a tool that can be helpful to researchers and authors by reducing the burdens of some tasks and improving efficiency in some contexts,” says the JAMA editorial, which Dr. Bibbins-Domingo co-wrote with Annette Flanagin, RN, MA, and Roy H. Perlis, MD, MSc. Flanagin is the executive managing editor and vice president of editorial operations for JAMA and the JAMA Network; Dr. Perlis is editor-in-chief of JAMA+ AI.
In a table, the JAMA editorial outlines some acceptable uses of AI. For example, using AI tools to check grammar or spelling is fine and requires no disclosure. But for research, literature searches or formatting data or graphic visualization of research data or results, the authors must fully describe the use in the methods section of the article (or the figure legend, for graphics), double-check for accuracy, and “take full responsibility for all content.”
These uses for AI tools in preparing manuscripts or other content are all listed as unacceptable:
- Presenting AI-generated data as if it were original research from non-AI sources.
- Drafting letters and online comments.
- Drafting opinion manuscripts.
- Clinical images and illustrations.
- Clinical video and audio.
- Peer review.
- Author response to peer reviewer and editor comments.
The use of AI tools for bibliographic references, meanwhile, is “discouraged,” says the editorial.
An accompanying box in the editorial adds that “AI, LLM and chatbots should not be used to generate or format references because these tools may cause problems with the accuracy and integrity of cited references; standard reference managers may be used instead. Check all references for completeness, accuracy and integrity.”
If an AI tool is used to help create, revise or format content, authors should include in their acknowledgments information about the name of the AI software platform, program or tool used, the version and extension numbers, the manufacturer, the dates of use, a brief description of how it was used and on what portions of the manuscript.
Despite the new guidance for authors on use of AI tools, JAMA’s editors have a positive view of how AI tools can help advance science and medicine, Dr. Bibbins-Domingo said.
“At JAMA, we're quite bullish on AI,” she said. “We know that it's going to be something that transforms how we practice medicine. It transforms how we do science, and it changes publishing. ... It makes publishing more accessible. It means more people for whom English is not their first language, who may not have published in the form that JAMA is used to evaluating papers, are able—that it’s more accessible to them because they can use these tools to really do that.”
Telltale signs of AI slop
But such uses must be disclosed, and the new guidance offers more detail on how to do that. According to research published in JAMA in January, of more than 100,000 manuscripts submitted to 13 JAMA Network journals between August 2023 and October 2025, only 3.3% declared use of AI.
“We just don't really believe that,” Dr. Bibbins-Domingo said. “It's just way too low,” noting that disclosing use of AI for translation—which is permitted—pushed the figure up to 6%.
“Probably the best indicator that AI is being used out of proportion is in our letters to the editor. These are short letters that you could write in and say: ‘That article, I didn't like this about it.’ Or, you know, that they didn't use the right methods, or that's totally wrong for whatever reason. We've seen a massive spike in those types of things,” she noted.
“And these are particularly easy things to feed an article into AI and to say: ‘Give me 300 words critiquing this study.’ What are the telltale signs? The telltale signs are the very formulaic writing. It has almost exactly the same paragraph that starts it. It has three points to support that paragraph, and it doesn't say much.”
Another telltale sign is that “all of a sudden people who have never commented on any article—ever—are commenting on multiple different types of articles,” Dr. Bibbins-Domingo said. “So you might be writing us about an article we did on gastroparesis, and hip fracture, and sepsis ... and you're going to comment on all of those, which would be an unusual pattern.”
Other medical journals are adopting similar editorial policies on use and disclosure of AI tools, Dr. Bibbins-Domingo said, citing The BMJ and the International Committee of Medical Journal Editors.
Ultimately, physicians and other readers of JAMA and the JAMA Network should take solace in the journals’ stringent editorial processes.
“In this age of AI, the value that our editorial teams, our peer reviewers, and the authors who get published in JAMA put in is actually at a premium even more than it has been in the past,” Dr. Bibbins-Domingo said. “We spend a lot of time and care at JAMA so that in the end—even when authors are declaring whatever they're declaring regarding their use of AI—all of our processes serve as checks and balances to make sure that we know who's actually submitting an article, that we know that the references are all real. ... The reader of JAMA should be reassured that the entire process of what we do in editorial and publishing at JAMA is to make sure you can trust what you read.”
From AI implementation to digital health adoption and EHR usability, the AMA is fighting to make technology work for physicians, ensuring that it is an asset to doctors. That includes the AMA Center for Digital Health and AI, which works to ensure physicians help shape how AI is developed, implemented and regulated across the healthcare system, with patient safety and physician-led care remaining at the center of those efforts.
