Mathematician Publishes AI-Generated Fake Article in Predatory Journal
A mathematics professor at Universitat de València used ChatGPT to write a deliberately nonsensical article on obstetrics that was accepted and published by a predatory journal within weeks, exposing how AI-generated misinformation can bypass peer review in fraudulent academic publishing.
What happened
In October 2025, Pascual Diago, a mathematics education professor at Universitat de València, submitted a deliberately nonsensical article to the Clinical Journal of Obstetrics and Gynecology using ChatGPT's free version. The article, titled 'Paradojas obstétricas y ecuaciones didácticas: El impacto de la enseñanza de las matemáticas en el parto y más allá', discussed mathematical concepts applied to pregnancy and childbirth, a field outside Diago's expertise.
Diago included explicit markers of fraud throughout the paper: fabricated author names that were wordplay on English terms associated with deception (Cheatillo, Sneakydez, Trickón, Sneakarez), non-existent references, and four citations to studies by a researcher identified as Jamshidkhan Chamani from Iran's Islamic Azad University. The journal accepted the article within weeks and requested additional citations to Chamani's work.
Reported impact
The journal published the false article without rejection. Diago later received an invitation to speak at conferences based on the published work. The incident demonstrates how predatory journals publish content without meaningful peer review, regardless of scientific validity or author credentials.
Index analysis
Diago stated his intention was both partly a joke and partly a denunciation of predatory publishing practices. He described the article as 'incoherent', saying he could not understand its own content. The experiment exposed the vulnerability of the academic publishing system to AI-generated misinformation.
The professor reported that the journal editorial staff made limited substantive demands, requesting mainly additional citations and payment for publication. He stated the only consistent request was payment, and he communicated with four supposed editorial staff members.
Governance controls
Predatory journals operate without quality control mechanisms. The list of identified predatory journals exceeded 20,000 as of February 2026, up from 15,000 five years prior, according to the specialised monitoring firm Cabells.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.