Science advances by building on what came before — a chain of citations that links each new claim to the evidence beneath it. A joint study from Cornell and UCLA has found that chain compromised at scale: nearly 147,000 fabricated references, conjured by AI systems that generate the appearance of knowledge without its substance, now inhabit four of the world's major scientific repositories. The rupture is not the work of a few bad actors but a diffuse consequence of a tool adopted widely and verified rarely, arriving at a moment when the boundary between assistance and invention has grown dang
CNET
· May 21, 2026
· technology
· 2 min de lectura
High
Verificación: Verificado
Study Finds 146,900 AI-Generated Fake Citations Polluting Scientific Research
◉
Technology
8894 artículos relacionados
Google Launches Gemini Enterprise for Legal Services
Google News · 100% coincidencia
Nvidia Q3 earnings loom as options traders brace for 7.69% volatility …
The Economic Times · 99% coincidencia
Apple's Siri AI Leverages Google Gemini While Plotting Deep App Integr…
digitimes · 99% coincidencia
Trump's Q1 2026 Stock Trades Reveal $220M-$750M in Big Tech Bets
News18 · 99% coincidencia
Nvidia Shield Users Report Dark Video, CEC Glitches, Overheating Issue…
bgr.com · 99% coincidencia
Cobertura Relacionada
The New York Times · Aug 27
China's AI Optimism Outpaces American Skepticism
Chinese citizens show significantly higher optimism about artificial intelligence compared to Americans, reflecting dive…
The Age · Aug 27 Apple's camera AirPods signal new era of wearable AI—but regulation must catch upApple is reportedly developing camera-equipped AirPods, escalating the wearable camera debate beyond smart glasses. The …
Yahoo News Canada · Aug 27 AI-assisted surgery removes brain tumor in world first for British patientBritish surgeons used AI to successfully remove a brain tumor from patient Rhys Hibbert, marking the first such procedur…
CNBC · Aug 27 Z.ai shares surge 8% on homegrown chip AI model launchChinese AI company Z.ai released GLM-5.3-Flash, claiming it runs entirely on domestically-made semiconductors, driving a…
Sesgo y Encuadre
⚖
CNET reports on a Cornell-UCLA study documenting AI-generated fake citations in scientific papers with balanced framing, though emphasizes problem severity over nuanced context.
Bias Direction
center-left
Confidence
MED
Problem-focused alarm framing that emphasizes the threat to scientific integrity and public trust, using dramatic language ('disturbing,' 'erode faith') while presenting the issue as widespread and systemic rather than isolated.
disturbing
threatening the integrity
erode faith
sloppy science
unscrupulous researchers
hallucinations
fabricated
Researchers or institutions defending their citation practices or explaining accidental errors
Database administrators' efforts to detect or prevent fake citations
Perspective from AI developers on safeguards being implemented
Quantification of how many papers are actually affected (percentage of 2.5M)
Discussion of existing peer review mechanisms that should catch such errors
Voices from researchers who successfully use AI tools responsibly
Impacto Geopolítico
🌎
AI-generated fake citations in scientific papers threaten global research integrity and institutional trust, with geopolitical implications for technology leadership and scientific credibility.
Escalation Risk
moderate
Confidence
HIGH
Global scientific community
United States
European Union
China
India
All nations dependent on peer-reviewed research
Nations with strong AI regulation (EU) gain credibility advantage over permissive markets (US, China). Scientific leadership shifts toward institutions implementing rigorous verification protocols. Developing nations' research credibility may suffer disproportionately if they lack resources for AI-detection infrastructure.
Similar to the replication crisis in psychology (2010s) and scientific fraud scandals (Hwang Woo-suk cloning hoax, 2005), which eroded public trust and required institutional reforms. This AI-driven version spreads faster and wider.
Lente Económico
📈
146,900 AI-generated fake citations discovered in scientific papers threaten research integrity, with potential economic costs through reduced innovation efficiency, increased verification expenses, and erosion of trust in knowledge-dependent sectors.
Market Signal
bearish
Confidence
HIGH
Higher Education & Research
Publishing & Academic Journals
Pharmaceutical & Biotech
Technology & Innovation
Professional Services
Quality Assurance & Compliance
Consumers face indirect costs through delayed medical treatments, ineffective products, and higher prices as companies must invest in additional verification processes. Trust in research-backed claims (health products, technology, safety standards) diminishes, increasing consumer skepticism and decision-making friction.
Likely regulatory responses include: mandatory AI disclosure in academic publishing, stricter citation verification requirements, institutional liability frameworks, potential licensing requirements for AI tools in research, and increased funding for peer-review infrastructure. Academic institutions may face compliance costs and reputational liability.