&w=1200&q=75)
Surge in AI-Generated Citations Threatens Scientific Integrity
Scientists Warn AI Slop Is Wreaking Havoc in the Research World
Image: Cnet
A study by researchers from Cornell University and UCLA reveals that approximately 146,900 AI-generated fake citations have infiltrated scientific papers across major research databases. This trend, linked to the rise of large language models like ChatGPT, poses a serious threat to the credibility of scientific research, undermining trust in academic publications.
- 01The study analyzed 111 million references from 2.5 million scientific papers, revealing a significant increase in unmatched citations since the advent of large language models.
- 02Researchers found that fake citations are widespread, indicating many scholars are using AI-generated references without proper verification.
- 03Usha Haley, a professor at Wichita State University, expressed concern that fake citations are eroding trust in the scholarly record essential for peer review.
- 04The four databases affected include arXiv, bioRxiv, SSRN, and PubMed Central, which are critical for the dissemination of scientific knowledge.
- 05In response to the issue, arXiv has announced a ban on authors who submit papers containing hallucinated citations or unverified AI content.
Advertisement
In-Article Ad
A recent study conducted by researchers from Cornell University and UCLA has uncovered a troubling trend in scientific literature: the presence of 146,900 AI-generated fake citations in papers across four major research databases—arXiv, bioRxiv, SSRN, and PubMed Central. This issue stems from the limitations of large language models like ChatGPT, which can generate plausible yet incorrect information, a phenomenon known as hallucination. The research team analyzed 111 million references from 2.5 million scientific papers, finding a marked increase in unmatched citations post-2023, coinciding with the widespread adoption of AI tools. Usha Haley, a management professor at Wichita State University, highlighted the dangers of this trend, stating that it undermines the trust essential for peer review and scholarly integrity. The findings raise alarms within academia, particularly among early career scholars, regarding the reliability of research. In response, arXiv has taken proactive measures by banning authors who submit works with unverified AI content, emphasizing the need to maintain the integrity of scientific discourse. As the research community grapples with this challenge, the dilution of credible scientific information poses risks to innovation and public trust.
Advertisement
In-Article Ad
The rise of AI-generated citations could lead to a decline in the quality and trustworthiness of scientific research, affecting researchers, institutions, and the integrity of academic discourse.
Advertisement
In-Article Ad
Reader Poll
Do you believe AI tools should be used for academic research?
Connecting to poll...
More about Cornell University
Cornell University Develops Self-Organizing Robot Swarm Using Physics
Interesting Engineering • May 21, 2026
Cornell's Center for Advanced Technology Promotes Life Sciences Innovation
Cornell Chronicle • May 20, 2026
Tribute to Albert ‘Al’ George: Pioneer in Experiential Learning and Engineering Education
Cornell Chronicle • May 19, 2026
Read the original article
Visit the source for the complete story.