In the long tradition of human error finding new vessels, artificial intelligence has introduced a novel form of scholarly corruption: citations that look real, sound real, and point to nothing at all. A Columbia University researcher, embarrassed by a fabricated reference in his own work, turned that moment into a sweeping investigation — and found that across nearly 2.5 million biomedical papers, the rate of invented citations has grown twelvefold in just three years. What troubles observers most is not the scale of the deception, but where it is landing: inside the clinical guidelines that
MedPage Today
· May 07, 2026
· technology
· 4 min de lectura
High
Verificación: Verificado
AI-Fabricated Citations Surge 12-Fold in Biomedical Research, Study Warns
Patients may receive clinical care based on guidelines citing non-existent research, potentially compromising treatment safety and efficacy.
◉
Technology
8676 artículos relacionados
Trump's Q1 2026 Stock Trades Reveal $220M-$750M in Big Tech Bets
News18 · 100% coincidencia
WhatsApp's Liquid Glass redesign rolls out to Mac with unified interfa…
Digital Trends · 99% coincidencia
Apple Sues OpenAI for Trade Secret Theft by Former Employees
CNET · 99% coincidencia
Apple sues OpenAI for allegedly stealing trade secrets through former …
The Straits Times · 99% coincidencia
US Judge Allows Musk's Antitrust Lawsuit Against Apple, OpenAI to Proc…
Times of India · 99% coincidencia
Cobertura Relacionada
infoq.com · Aug 23
Researcher Exposes CPU Memory Isolation Flaw via DRAM Controller Manipulation
Security researcher Christopher Domas unveiled a hardware exploit that bypasses CPU privilege boundaries by manipulating…
Memeburn · Aug 23 Fairphone Gen 6+ Brings True Repairability to US Market at $649Fairphone launches its first US smartphone at $649 with 12 user-replaceable parts, removable battery, and six years of s…
The Times of India · Aug 23 Learning to Code Still Matters—Just in Different Ways, Microsoft SaysMicrosoft argues coding remains essential despite AI generating 20-95% of code at major tech firms, shifting the skill f…
Al Jazeera · Aug 23 Chinese humanoid robot shatters Bolt's 100m record at Beijing gamesA Chinese humanoid robot named Tianzhuo ran 100m in 9.39 seconds at the World Humanoid Robot Games, surpassing Usain Bol…
Sesgo y Encuadre
⚖
Article presents factual reporting on a legitimate research study about AI-fabricated citations with minimal apparent bias, though framing emphasizes the problem's severity.
Bias Direction
center
Confidence
MED
Problem-focused narrative emphasizing crisis escalation through dramatic statistics (12-fold increase, exponential growth trajectory). The 'tip of the iceberg' quote amplifies concern beyond measured findings.
surge
fabricated
hallucinating
deeply embarrassed
tip of the iceberg
increasing problem
AI tool developers' responses or mitigation efforts
Counterarguments about detection effectiveness improving
Perspective from researchers defending AI tool utility
Discussion of false positive rates in detection methodology
Context on historical citation error rates pre-AI
Impacto Geopolítico
🌎
AI-generated fabricated citations in biomedical research surge 12-fold since 2023, threatening scientific integrity and potentially undermining global health research credibility and international collaboration standards.
Escalation Risk
moderate
Confidence
HIGH
Global scientific community
United States
European Union
China
India
All nations dependent on biomedical research
Erosion of Western scientific institutional authority as AI-driven misinformation undermines peer review systems. Nations with robust research verification infrastructure (US, EU) may gain relative advantage over those with weaker oversight. Potential shift toward stricter international research standards favoring established institutions.
Similar to the reproducibility crisis in psychology (2010s) and stem cell research fraud (Hwang Woo-suk, 2005), but with AI acceleration making detection and prevention more challenging. Parallels Cold War scientific competition where falsified data occasionally went undetected.
Lente Económico
📈
AI-generated fabricated citations in biomedical research surged 12-fold since 2023, threatening research integrity and potentially undermining pharmaceutical development, clinical trials, and healthcare policy decisions based on flawed evidence.
Market Signal
bearish
Confidence
HIGH
Pharmaceutical and biotechnology
Healthcare and medical devices
Academic publishing and research
Regulatory affairs and compliance
Healthcare insurance and payers
Clinical trial services
Patients face increased risk from medical treatments and drugs developed based on compromised research; healthcare costs may rise as flawed studies lead to ineffective therapies or adverse outcomes requiring additional treatment.
Regulatory bodies (FDA, EMA) may mandate stricter citation verification protocols; journals may require AI disclosure and enhanced peer review; potential legislation requiring AI-generated content labeling; increased compliance costs for publishers and research institutions; possible retractions affecting drug approvals.