In 2019, the US National Institute of Standards and Technology (NIST) tested 189 facial-recognition algorithms from 99 developers on millions of images. NIST did not ask only whether a machine identified the correct person. It also measured how the machine failed across age, sex, and demographic groups. The results showed differences across most algorithms. In many one-to-one comparisons, some Asian and African American faces were falsely matched more often than white faces, with certain algorithms showing differences of ten to one hundred times. In one-to-many searches, African American women also had higher false-match rates in many algorithms.
In 2019, the US National Institute of Standards and Technology (NIST) tested 189 facial-recognition algorithms from 99 developers on millions of images. NIST did not ask only whether a machine identified the correct person. It also measured how the machine failed across age, sex, and demographic groups. The results showed differences across most algorithms. In many one-to-one comparisons, some Asian and African American faces were falsely matched more often than white faces, with certain algorithms showing differences of ten to one hundred times. In one-to-many searches, African American women also had higher false-match rates in many algorithms.
In 2019, the US National Institute of Standards and Technology (NIST) tested 189 facial-recognition algorithms from 99 developers on millions of images. NIST did not ask only whether a machine identified the correct person. It also measured how the machine failed across age, sex, and demographic groups. The results showed differences across most algorithms. In many one-to-one comparisons, some Asian and African American faces were falsely matched more often than white faces, with certain algorithms showing differences of ten to one hundred times. In one-to-many searches, African American women also had higher false-match rates in many algorithms.
In 2019, the US National Institute of Standards and Technology (NIST) tested 189 facial-recognition algorithms from 99 developers on millions of images. NIST did not ask only whether a machine identified the correct person. It also measured how the machine failed across age, sex, and demographic groups. The results showed differences across most algorithms. In many one-to-one comparisons, some Asian and African American faces were falsely matched more often than white faces, with certain algorithms showing differences of ten to one hundred times. In one-to-many searches, African American women also had higher false-match rates in many algorithms.
In 2019, the US National Institute of Standards and Technology (NIST) tested 189 facial-recognition algorithms from 99 developers on millions of images. NIST did not ask only whether a machine identified the correct person. It also measured how the machine failed across age, sex, and demographic groups. The results showed differences across most algorithms. In many one-to-one comparisons, some Asian and African American faces were falsely matched more often than white faces, with certain algorithms showing differences of ten to one hundred times. In one-to-many searches, African American women also had higher false-match rates in many algorithms.