Augmented eye clinic shows why AI-native systems work where standalone algorithms don't
Researchers deployed an AI-native multi-agent system, AI-TEC, across the ophthalmic care pathway at Beijing Tsinghua Changgung Hospital to handle patient intake, test recommendations, retinal imaging analysis, and follow-up. Although fine-tuning with expert-verified images significantly boosted diagnostic accuracy for conditions like glaucoma and diabetic retinopathy, clinician adoption plummeted from 25.7% to 3.8% when the system proved slow and click-heavy before rebounding after an interface redesign. For medical and health administration educators, this demonstrates that real-world clinical AI utility depends as much on usability and workflow integration as it does on algorithmic perform
Read the original at medicalxpress.com Generate teaching or study materials
More in Medicine
- How Roche drives AI outcomes in a changing worldCIO.com · September 16, 2026
- New portable blood testing device could allow patient health to be monitored at home through an appMedical Xpress · September 16, 2026
- From 7,000 drugs to one active against resistant strains: AI advances pneumonia drug repurposingMedical Xpress · September 16, 2026
- AI Isn’t a Therapist, but Many Teens Are Treating It Like OneMedCity News · September 16, 2026
- Study finds AI scribes save doctors' time and may ease administrative burdenMedical Xpress · September 16, 2026