A pilot feasibility case series evaluated Ray-Ban Meta AI eyeglasses (Generation 2) for accuracy across tasks relevant to low/no vision users: single and multiple object identification, color discrimination, directionality, object counting, reading (labels, handwriting, children's books), and US paper/coin money identification.
Object identification was highly accurate (99%; 699/700 trials). Performance was strong for handwriting (88%) and children's books (93%), moderate for directionality (83%) and paper money (91%), but poor for color discrimination (64%), standard text reading (59%), object counting (50%), and coin identification (2%).
- Only 6 sighted study authors served as participants — no patients with actual low or no vision were tested. - Very small, non-representative sample limits generalizability. - Controlled white tabletop/background used for most tasks, which may inflate real-world performance.
Ray-Ban Meta AI eyeglasses show promise for object identification and handwriting/book reading in low-vision contexts, but coin identification and object counting are currently unreliable. Counsel patients on these specific gaps before recommending the device.
Explore related topics
