An audit of 2.5 million biomedical papers reveals fabricated references have surged more than twelvefold since 2023, likely driven by language model use. The fake citations are nearly impossible to detect and threaten the integrity of clinical guidelines.
Researchers from Columbia University and partner institutions discovered that AI-hallucinated citations are spreading through the biomedical literature at an alarming rate. The fabricated references match paper topics, follow proper formatting conventions, and blend seamlessly into legitimate research.
The fake citations pose a particular danger because they often appear in papers that inform clinical practice guidelines—affecting real-world medical decision-making. Language models trained on existing research can generate plausible-sounding but entirely fictional sources that researchers may unknowingly cite further, amplifying the contamination.
Key concerns:
- 98 percent of affected papers remain unaddressed by publishers
- Detection requires manual verification of each citation
- Risk of cascading citations spreading false information
The finding highlights the tension between AI's integration into research workflows and quality control mechanisms designed for human-generated content. Publishers face mounting pressure to implement automated citation verification systems before hallucinated references become endemic to the scientific record.
U.S. federal prosecutors have unsealed charges against three Russian nationals accused of operating a bulletproof hosting service that supported ransomware gangs responsible for over $62 million in damages worldwide.
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) warned that attackers are actively exploiting three vulnerabilities in Internet-exposed on-premises SharePoint Server instances. Organizations running affected versions must patch immediately.
Tailscale disclosed a critical vulnerability in its SSH implementation that allowed attackers to gain root access through insecure argument handling. The flaw has been patched in recent versions.
A new study found that social media platforms referred over 5.7 million visits to nonconsensual deepfake pornography sites between December 2025 and March 2026, with YouTube and X accounting for the majority of traffic.