Artificial intelligence and measurement: what it really changes for opticians

"Artificial intelligence" has become such a common marketing claim that it deserves clarifying: concretely, what does AI actually do — and not do — in a centering measurement tool like IO Tech?

What AI does well

• Facial and eye detection: precisely locating facial landmarks (eyes, nose bridge, frame position) in every image, including across very different faces, face shapes, and skin tones.
• Cross-checking multiple protocols: automatically comparing distance, near, and screen-distance measurements to spot an inconsistency, rather than relying on a single reading.
• Real-time quality control: checking that posture (head tilt, alignment) is correct before validating a measurement, and alerting the optician if it isn't.

What AI doesn't replace

AI doesn't replace the optician's professional judgment: choosing the lens, adjusting the frame, and the wearer's overall clinical assessment remain human decisions. Nor can it infer information that the measurement itself doesn't contain — if the frame is poorly fitted on the face at the moment the photo is taken, no algorithm can correct that bias afterward. AI makes the measurement more reliable; it doesn't substitute for an eye exam or for the professional's expertise.

Why cross-checking matters more than raw precision

A single algorithm, however precise, can get a particular image wrong (a reflection on a lens, a stray strand of hair, unusual lighting). Reliability comes less from the precision of one model than from comparing several independent measurements: if distance and near vision give results consistent with the convergence shift expected at those distances, the measurement is validated; if not, the tool asks for a new reading rather than delivering an unreliable number.