Facial age estimation with fallback to ID verification. Includes audit-backed fairness metrics and governed update process; privacy-preserving design with opt-out controls.
Persona orchestrates flows using selfie-based estimation first, escalating to document or NFC verification only when age proximity or quality issues trigger risk-based fallback.
Supports user opt-out and clear user interface, strong performance with First Nations youth in school trials, demographic audit trails supported. MAE ranged from 0.86 to 3.31 across different cohorts; strong at 13+ thresholds. Accuracy varied more widely for under-13 users.
