Uses a neural network to estimate a person's age by analyzing facial features from an image, with no need for physical documents or human intervention. It is designed to be privacy-preserving, immediately deleting all images after estimation and not storing or reusing them. The technology is accurate across gender and skin tone, with a mean absolute error of 1.36 years for 6-12 year olds and 1.52 years for 13-19 year olds, and high true positive rates for verifying underage individuals. It is built in accordance with GDPR principles, supports data minimisation, and does not identify individuals or track user behavior. The system is certified by regulators such as the German KJM and the US NIST, and is used to prevent underage access to age-restricted content and goods while promoting social inclusion for those without identity documents.
