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 document checks or human intervention. The technology adheres to GDPR principles, emphasizing privacy by design and data minimization. It is designed to be unbiased across gender and skin tone, with a True Positive Rate (TPR) of 99.3% for individuals aged 13 to 17 and 99.0% for those aged 6 to 12. The model has a mean absolute error (MAE) of 1.1 years for ages 13 to 17 and 1.3 years for ages 6 to 12. The white paper details the ethical considerations in data collection, ensuring compliance with GDPR standards. It discusses the technology's applications in various sectors, including retail and online services, and its role in safeguarding children by preventing access to age-inappropriate content. The document also outlines the performance evaluation by NIST and future improvements in accuracy.
