Removing bias within facial verification AI with more diverse data

Human in The Loop, a strategic partner of UNISSEY, describes the activity of the startup and highlights key points of their partnership. In this article, you will get details about the work done to make the facial biometric solution developed by UNISSEY, universal, without discrimination.

In its quest to offer a universal and accessible facial biometric solution, UNISSEY has partnered with HITL (Human in the Loop), which is responsible for providing as much diverse data as possible to test and strengthen its solution. In other words, HITL is full of biometric data from people all over the world, of all races and ages (in compliance with RGPD standards, having obtained their consent beforehand) that can be used to strengthen the algorithms and avoid technological bias.

What are technological biases? Indeed, certain problems can occur when presenting certain profiles. For example, a 50 year old man or woman who presents a driver’s license with a picture of their adolescence can be systematically rejected by the solution if it has not been trained to deal with this type of case. This is why a partner like HITL is indispensable in the development of such technology.

UNISSEY’s ambition is to offer the best service to EVERYONE, without any discrimination. This is a real work that is done in this sense and it would not be possible without partners like HITL.

The rest in the article is available here →