Incode states that with On-Device Age Estimation a user's selfie is never sent anywhere, and only minimal session metadata goes to its servers.
How Incode On-Device Age Estimation Remains Accurate, Secure, and Private ↗
What happened
- Incode states it reduced the on-device age model size by 10x compared with its traditional server-side model.
- Incode reports 1.08–2.08 years mean absolute error for its on-device age model across adult and youth groups, versus 0.95–1.87 for its server-side model.
- Incode states it does not collect users' IP addresses or fingerprint devices for privacy-focused use cases.
Why it matters to Signicat
- Signicat's age verification and privacy positioning faces a competitor claim of on-device biometric processing; Signicat's public materials in this input do not describe an equivalent on-device age model.
- The MAE and 10x figures are vendor self-reported without published test methodology and are not independently verified.
Evidence from the source (5 quotes)
- “we reduced the size of our On-Device Age Estimation model by 10x compared to our traditional server-side model.” source ↗
- “our on-age model has achieved 1.08–2.08 years mean absolute error (MAE) for both adult and youth age groups, compared to 0.95-1.87 MAE for our server-side model.” source ↗
- “For privacy-focused use cases, we deliberately chose not to collect users' IP addresses or fingerprint their device.” source ↗
- “But for On-Device Age Estimation, we went further.” source ↗
- “our on-age model has achieved 1.08–2.08 years mean absolute error (MAE)” source ↗
Age verificationData sovereignty privacyDocument IDV biometricsFraud deepfake injectionProduct launchThought leadershipGlobal
4 of 4 facts verified against the source · confidence medium · claude-haiku-5-5