Radiologists may misidentify AI images
CT and MRI images were harder to distinguish than radiographs and ultrasound, while subspecialty expertise was associated with greater accuracy.
By
Andrea Surnit
August 26, 2026
Clinical Scorecard: Radiologists may misidentify AI images
At a Glance
Category Detail
Condition Artificial Intelligence in Radiology Key Mechanisms Identification of AI-generated vs. real radiological images Target Population Radiologists Care Setting Clinical Radiology
Key Highlights
Radiologists correctly identified 75% of AI-generated images. Accuracy varied by imaging modality: CT (70%), MRI (77%), Ultrasound (88%), Radiographs (91%). Relevant subspecialty expertise improved classification accuracy. Confidence in classification correlated with correct identification. Study limitations included online presentation of images and potential selection bias. Guideline-Based Recommendations
Diagnosis
Assess classification accuracy of AI-generated images across modalities. Management
Consider subspecialty expertise when evaluating AI-generated images. Monitoring & Follow-up
Monitor radiologist confidence levels in image classification. Risks
Potential misidentification of AI-generated images may impact diagnostic accuracy. Patient & Prescribing Data
Not specified; study focused on radiologists.
N/A
Clinical Best Practices
Incorporate training on AI-generated images in radiology education. Encourage radiologists to report confidence levels during image classification. Related Resources & Content
