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 is designed to be compliant with GDPR principles, ensuring privacy by design and data minimization. It does not identify individuals and only requires a facial image for age estimation. The paper details the True Positive Rates (TPR) for various age groups, demonstrating high accuracy across genders and skin tones. For instance, the TPR for 13 to 17 year olds is 99.91%, while for 6 to 12 year olds, it is 96.99%. The technology utilizes a neural network trained on facial images and birth dates, collected in accordance with GDPR guidelines. The paper also discusses the ethical responsibilities Yoti undertakes in developing this technology, including efforts to reduce bias and improve accuracy across diverse populations. Continuous improvements and compliance with regulatory standards are emphasized throughout the document.
