Most face recognition algorithms have more difficulty recognizing black and Asian people than white faces. This is the result of a study of 189 algorithms from large companies. There are also more different results between women than among men.
The study from the American National Institute of Standards and Technology looked at 189 algorithms from 99 different developers. This involves systems from large companies, such as Microsoft and Intel. The research looked mainly at empirical results but does not draw conclusions about the causes of the prejudices of the algorithms. The study looked at both matching a face with one specific photo, such as when using a biometric face scanner on a telephone, and looking at matching a photo from a large collection of photos, as is often the case with tracing. The researchers note that the differences between black and white people are particularly large in the first group. The results differ when matching one face with multiple faces. “Not all algorithms have that”false positives. Different algorithms produce different results, “they write.
The research shows that in some cases black and Asian men are incorrectly considered a match with a photo 100 times as often as white men. Other ethnic groups, such as native Americans and people with a Polynesian appearance, are highlighted significantly more often in the algorithms than white people. The algorithms also make a distinction between age and gender. Women are more often unjustly highlighted, just like the elderly and children. According to the researchers, this is certainly a danger with detection algorithms. “In a one-to-one comparison, a false negative is just an inconvenience that you can’t get into your phone, for example. But with a false positive compared with other images, an incorrect match may lead to further investigation, “write the researchers.
There has been a lot to do about facial recognition lately, especially by the police and other investigative services. Various politicians from CDA and D66, among others, want a temporary stop for this, something that is also endorsed by Justice and Security Minister Grapperhaus and the European Commission . It would then be about the use of face recognition for detecting criminals, but also by tech companies that build it into their products and services.