AI Misclassification
Cross-source consensus on AI Misclassification from 1 sources and 5 claims.
1 sources · 5 claims
Benefits
Risks & contraindications
Evidence quality
Highlighted claims
- The AI algorithm misclassified 9.5% of cases as referable diabetic retinopathy. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
- AI-assisted screening has potential but requires careful integration because of misclassification risk. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
- False positives may increase unnecessary referrals, burden health systems and cause patient anxiety. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
- False negatives are more dangerous because they can miss treatment referrals and increase preventable vision loss risk. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
- The AI model's high negative predictive value helped reduce false-negative screening failures. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study