In a sweeping crackdown on the use of artificial intelligence in scholarly writing, one of the most recognized journals in its field has introduced a stringent policy that bans any manuscript containing material generated by large language models (LLMs) like ChatGPT. The announcement, which was published on the journal's website earlier this week, marks one of the strongest stances taken by an academic publication since the widespread adoption of generative AI tools began in late 2022.
The journal—which specializes in interdisciplinary social sciences and has an impact factor above 10—will now require all authors to submit a signed declaration stating that no part of their manuscript was produced using an AI text generator. Editors and reviewers will also be trained to detect telltale signs of machine-written prose, such as unnatural sentence structures, repetitive phrasing, and an overreliance on certain transitional words.
Rising Concern Over AI in Academia
The decision comes amid a surge of AI-assisted submissions across multiple disciplines. In a 2025 survey by the Committee on Publication Ethics, 68 percent of journal editors reported encountering at least one submission they suspected was largely written by AI. The percentage was even higher for journals in computer science, engineering, and the humanities. Many editors described a growing frustration with having to vet manuscripts for machine-generated content instead of focusing on scientific merit.
Dr. Emily Hartfield, the journal's editor-in-chief, explained the rationale in an interview: "Our mission is to advance knowledge through rigorous, original human scholarship. Allowing AI to write papers undermines the very foundation of peer review and intellectual contribution. We cannot credit a machine as an author, yet the text it produces carries no accountability for accuracy or ethical compliance." The journal also emphasized that using AI for minor tasks like grammar checking or reference formatting remained acceptable, as those do not affect the core intellectual work. However, any use of AI to generate entire paragraphs, literature reviews, or data interpretation would be grounds for immediate rejection.
How the Policy Works
Authors submitting a new paper will be required to check a box during the online submission process attesting that all text is their own original work. They must also upload a separate statement describing any use of AI tools—even if only for editing or translation. Failure to disclose such use could result in a ban from submitting to the journal for three years. The journal plans to randomly audit a percentage of published papers each year using plagiarism detection software that can now identify AI fingerprints.
One challenge the journal acknowledges is that detection is not always reliable. Current tools like GPTZero and Originality.ai have false positive rates that can be as high as 10 percent. To mitigate this, the journal will rely on a combination of software screening and human review from seasoned editors who are familiar with the writing style of their field. "We don't want to punish scholars who write clearly or who use standard academic phrases," Hartfield added. "But we have to draw a line somewhere."
Broader Implications for Research
The crackdown has sparked debate across the academic community. Supporters argue that it restores integrity to a system already strained by predatory journals and pressure to publish. Opponents, however, point out that AI tools have legitimate uses—especially for non-native English speakers who rely on them to polish grammar and achieve clarity. "Blanket bans risk punishing those who most need assistance to level the playing field," said Professor Mei Lin, a linguist at the University of Toronto. "We should instead develop guidelines for transparent disclosure."
Several other journals are watching the development closely. The Journal of Medical Ethics has already announced a similar policy, while Nature and Science have issued statements requiring authors to disclose AI use but stopping short of an outright ban. The trend suggests that the academic publishing industry is entering a period of adjustment, as it tries to balance the efficiency gains of AI with the imperative of human-authored originality.
Historical Context and Future Outlook
The issue is not entirely new. In 2023, several preprint servers began flagging papers that appeared to be AI-generated after a flood of nonsensical submissions. By 2024, the U.S. Copyright Office had ruled that works wholly created by AI are not eligible for copyright, and the Supreme Court declined to hear a challenge to that position earlier this year. Meanwhile, lawsuits against OpenAI and other companies have raised questions about whether training on copyrighted academic papers constitutes infringement.
For researchers, the message is clear: as AI continues to evolve, the definition of authorship will need to be rethought. Some scholars advocate for a new category—"AI-assisted" work—where the machine serves as a tool but the human retains responsibility for every sentence. Others argue that any use of generative text is antithetical to the scientific method, which demands transparent and replicable steps that cannot be performed by a black-box algorithm.
The journal's move may accelerate efforts to develop robust detection standards and clearer editorial policies across disciplines. In the meantime, authors are advised to err on the side of caution. "If you wouldn't want your name on it, don't let a chatbot write it," Hartfield concluded. The new policy took effect immediately, and the first batch of submissions under the stricter rules will be reviewed within the coming months. Early reactions from the academic community suggest that this is only the beginning of a larger shift in how scholarly publishing grapples with the age of generative AI.
Source: Mashable News