US EditionPublished • loading... • Updated
Health Rounds: Fake X-Rays Created by AI Fool Radiologists and Even ...
A study of 17 radiologists and AI models across six countries found only 41% initially detected deepfake X-rays, with accuracy rising to 75% after being alerted to fakes.
- On March 24, 2026, a study published in Radiology found that neither experienced radiologists nor advanced large language models can reliably distinguish authentic X-rays from AI-generated deepfake images.
- Seventeen radiologists from 12 medical centers across six countries reviewed 264 X-ray images; only 41% spontaneously identified AI-generated scans despite clinical experience ranging from zero to 40 years.
- Even when informed synthetic images were present, radiologist accuracy reached only 75%, while four large language models—GPT-4o, GPT-5, Gemini 2.5 Pro, and Llama 4 Maverick—ranged from 57% to 85% accuracy.
- Dr. Mickael Tordjman of the Icahn School of Medicine at Mount Sinai warned that deepfake X-rays "creates a high-stakes vulnerability for fraudulent litigation if, for example, a fabricated fracture could be indistinguishable from a real one."
- Researchers are calling for advanced digital safeguards including invisible watermarks and cryptographic signatures attached at image capture to verify authenticity and prevent malicious exploitation of generative AI tools.
Insights by Ground AI
13 Articles
13 Articles
Deepfake X-Rays Sneak Past Radiologists and AI, Underscoring Abuse Potential
(MedPage Today) -- A majority of radiologists could not distinguish artificial x-rays -- deepfakes -- from real ones when they evaluated a mix of real and fake images, according to a study published today. Initially, only seven of 17 radiologists...
·New York, United States
Read Full Article
Deepfake X-rays can deceive radiologists and AI systems
Neither radiologists nor multimodal large language models (LLMs) are able to easily distinguish artificial intelligence (AI)-generated "deepfake" X-ray images from authentic ones, according to a study published today in Radiology, a journal of the Radiological Society of North America (RSNA).
·United States
Read Full ArticleCoverage Details
Total News Sources13
Leaning Left1Leaning Right1Center9Last UpdatedBias Distribution82% Center
Bias Distribution
- 82% of the sources are Center
82% Center
C 82%
Factuality
To view factuality data please Upgrade to Premium
