Amazon Facial Detection Software Identifies Dark-Skinned Females as Male
Amazon’s much-vaunted facial recognition software - currently the subject of a concerted effort to halt its production - regularly identifies dark-skinned females as males, it has emerged.
The facial-detection technology—which Amazon is marketing to law enforcement bodies—“often misidentifies women, particularly those with darker skin, according to researchers from MIT and the University of Toronto.”
According to an AP report, “privacy and civil rights advocates have called on Amazon to stop marketing its Rekognition service because of worries about discrimination against minorities.” (“Minorities” is the controlled media code word for nonwhite, as whites are the actual minority, a fact which that media always endeavors to hide).
The researchers said that in their tests, Amazon’s technology labeled darker-skinned women as men 31 percent of the time. Lighter-skinned women were misidentified 7 percent of the time. Darker-skinned men had a 1 percent error rate, while lighter-skinned men had none.
Matt Wood, general manager of artificial intelligence with Amazon’s cloud-computing unit, said the study uses a “facial analysis” and not “facial recognition” technology. Wood said facial analysis “can spot faces in videos or images and assign generic attributes such as wearing glasses; recognition is a different technique by which an individual face is matched to faces in videos and images.”
This is not the first time that Artificial Intelligence has had “problems” with the automatic analysis of dark-skinned people. In 2015, Google was the subject of much media controversy when its Google Photos app regularly identified black people as “gorillas.”
Photo app Flickr also ran into problems in 2015, when its auto-tagging system labelled images of black people with tags such as “ape” and “animal” as well as tagging pictures of concentration camps with “sport” or “jungle gym.”
The system used what Flickr described as “advanced image recognition technology” to automatically categorize photos into a number of broad groups.
In 2018, Google admitted that it had been unable to solve the auto-identification of black people as “gorillas,” and had therefore resorted to simply preventing Google Photos from ever labelling any image as a gorilla, chimpanzee, or monkey – even pictures of the primates themselves.
As revealed in Wired Magazine, researchers tested Google Photos using a collection of 40,000 images well-stocked with animals. It performed impressively at finding many creatures, including pandas and poodles. But the service reported “no results” for the search terms “gorilla,” “chimp,” “chimpanzee,” and “monkey.”
In a third test attempting to assess Google Photos’ view of people, WIRED also uploaded a collection of more than 10,000 images used in facial-recognition research.
The search term “African American” turned up only an image of grazing antelope. Typing “black man,” “black woman,” or “black person,” caused Google’s system to return black-and-white images of people, correctly sorted by gender, but not filtered by race.
The only search terms with results that appeared to select for people with darker skin tones were “afro” and “African,” although results were mixed, Wired reported.
No-one can of course choose the way they appear, and it is mean-spirited to mock anybody because of their physical appearance.
Nonetheless, the continual struggle to get Artificial Intelligence software to pretend that race does not exist, just like the liberal race-denial so prevalent elsewhere in society—lies at the core of the problem in the examples illustrated above.
For, if instead of denying that race exists and is a biological reality, the software developers admitted that race was real and programmed it into their calculations, the misidentifications would certainly not occur.
For example, the identification of dark-skinned females as males is of course due to the well-known African racial trait of androgyny—the state where there is no clear distinction between male and female in facial appearance.
This racial trait is most clear in Africans of unmixed racial origin, such as these two early 20th Century French postcards from Africa reveal. On the left, a female from the Mogandi tribe in the Congo, and on the right, a male from the Bassoundi tribe, also in the Congo.
If liberals did not deny the reality of race, then software identification errors would not occur. In fact, if liberals did not deny the reality of race, then many problems facing the entire globe would not occur.