Monday , June 21 2021

Amazonian face-to-face technology shows sex, racial prejudice, says researchers

The facial-detective technology that Amazon is law enforcement, often rarely informs women, especially those with darker skin, according to researchers from MIT and the University of Toronto. Private and civil rights defenders called Amazon to Stop marketing its Recognition Due to concerns about discrimination against minorities.

Some Amazon investors have also requested that the company cease to fear that it is causing victims to Amazon through trials.

The researchers said that in their attempts, Amazon's technology tagged darker women as men 31% of the time. Lightweight women women have been misleading by 7% of the time. Dark-haired men had a 1 percent error rate, while lighter skin people had no one.

Artificial intelligence can imitate the prayers of its human creatures as they go into the daily life. The new study, released late on Thursday, warns of the potential of abuse and threats to privacy and civil liberties of facial detection technology.

Amazonian Protest
Demands hold images from Amazon CEO Jeff Bezos near their faces during a Halloween-based protest at Amazon headquarters over the system recognition system system, "Recognition", on October 31, 2018 in Seattle.

Elaine Thompson / AP

Matt Wood, chief of a director of artificial intelligence with the Amazon cloud cloud unit, said the study uses "easy analysis" and not "face recognition" technology. Wood said face analysis "can see faces in videos or images and assign general attributes such as carrying glasses; recognition is a different technique with which an individual face is equal to faces in videos and images."

On Friday, a post in the Middle Web, MIT Media Lab Joy Buolamwini's researcher responded that companies have to control all the systems that analyze human faces for justification.

"If you are selling one system that has been proven to be pressing human faces, it is uncertain that your other face-products are also free," she wrote.

The reaction of Amazon shows that it does not seriously emphasize the "serious concerns raised from this study," said Jacob Snow, a lawyer with the United States Civil Liberty Union.

Buolamwini and Inioluwa Deborah Raji from the University of Toronto said they had studied Amazon technology because the company sold it into law. Raji's LinkedIn contract says she is currently researching a lie for artificial intelligence at Google, who competes with Amazon in offering computer services.

Buolamwini and Raji say Microsoft and IBM have improved their face-to-face technology, because researchers have discovered similar problems in May 2017. Their second study, which included Amazon, was made in August 2018. Their paper will be presented on Monday at an artificial conference in Honolulu.

Wood said Amazon had updated its technology since the study and made its own analysis with "null fake positive matches".

The Amazon website recommends a resumption to assist the Sheriff Office of the District in Washington to accelerate how long it has identified suspicions of hundreds of thousands of photo maps.

Source link