IDV firm spins out facial age estimation startup making bold accuracy claims

A new entrant to the growing age assurance field is making some bold claims about the accuracy of its technology with its launch announcement claiming Agemin “outperforms all known competitors on the market.”

In addition to facial age estimation the company offers email age estimation and content age monitoring.

Agemin is a subsidiary of Bynn Intelligence, a San Francisco-based company that was incorporated in 2024 and acquired Swedish company Dokumentbedrägeri AB in July 2025, giving it access to the document forensic technology that Dokumentbedrägeri AB has been developing since 2003 (including the PDFChecker platform).

Agemin’s release claims its new model has achieved a Mean Absolute Error (MAE) of 0.9 years for the age group of 9-21 and 1.8 years for individuals 21 and older. The firm says it plans to submit the model to the National Institute of Standards and Technology’s (NIST) Face Analysis Technology Evaluation – Age Estimation & Verification (FATE/AEV) program for independent evaluation in the coming month.

NIST assesses age estimation based on Challenge 25 and Child Online Safety applications, and calculates MAE for three applications based on estimates of people between 18 and 30 years old. In the latest update, Innovatrics scored the lowest MAEs in the latter applications, for office scenarios (2.52), border photos (2.748)  and mugshots (2.389), and Yoti’s MAE for Child Online Safety was just over two years. Across subjects from different geographies (a proxy for skin tone), Regula had the lowest MAEs, ranging from 1.62 to 3.3.

“Our new model represents a significant advancement in biometric technology, delivering unprecedented accuracy tailored to the nuances of facial development in children and young adults,” says Mikael Hedlöf, CEO of Bynn Intelligence. The new model focuses on better performance specifically for younger demographics, and supports real-time deployment via APIs for easy integration.

“Frictionless age estimation is key to widespread adoption,” Hedlöf says. “Traditional methods like ID uploads create barriers that deter users and platforms alike. By making verification seamless and non-intrusive, we empower social media, gaming, and e-commerce players to implement robust child safety measures without compromising user experience or privacy.”

The company points to growing regulatory demands, such as those under COPPA, the UK’s Online Safety Act (OSA) and the EU’s Digital Services Act (DSA), as a key driver of demand for age verification and age estimation tools.

Per the release, Agemin’s new model will be available to enterprise partners in Q4 2025, pending validation from NIST.

Article Topics

age verification  |  Agemin  |  biometric age estimation  |  biometrics  |  Bynn Intelligence  |  face biometrics  |  facial age estimation (FAE)