by Rebecca Burgess | Jul 9, 2026
Some attacks are possible, and certified systems treat this as a robustness problem to be engineered against, not ignored. Liveness detection checks that the camera is looking at a real, present person rather than a photo, a screen replay or a mask, and is tested...
by Rebecca Burgess | Jul 9, 2026
Bias risk is real and must be managed, not denied. Facial recognition systems – not estimation systems – have historically been studied and did not perform as well for some groups, including people with darker skin tones and people whose appearance is affected...
by Rebecca Burgess | Jul 9, 2026
A margin of error is inherent in facial age estimation – it is a feature not a fatal flaw. It is designed into how the technology is used, exactly as it is offline. Shops apply policies such as “Challenge 25” because staff cannot judge 18 precisely, anyone who looks...
by Rebecca Burgess | Jul 9, 2026
The legal answer can depend on how a system is implemented, and we do not claim it is never biometric processing. What matters are the safeguards: whether the image is processed transiently, whether any biometric template is created, whether the system is...
by Rebecca Burgess | Jul 9, 2026
No, and the difference matters. Facial recognition tries to identify who someone is, by matching a face against a stored template or watchlist. Facial age estimation estimates an age or age range and never asks who you are. A compliant deployment will not create...
