Summary
This white paper presents Yoti's facial age estimation technology, which accurately determines a person's age from an image of their face without requiring physical documents or human intervention. The technology adheres to GDPR principles, ensuring privacy by design and data minimization. It distinguishes itself from facial recognition by not storing or matching images against a database, deleting all images immediately after age estimation. The document details the True Positive Rates (TPR) for various age groups, indicating high accuracy across genders and skin tones. For instance, the TPR for 13 to 17 year olds is 99.2%, while for 6 to 12 year olds, it is 99.5%. The paper also discusses ethical considerations, data collection practices, and the technology's application in safeguarding children online. Additionally, it outlines the deployment of facial age estimation in retail and law enforcement, emphasizing its effectiveness in age-restricted environments. Finally, it includes performance evaluations and comparisons from NIST assessments.
