Facial recognition software tests in the United States shows it is biased against ethnic minorities. Algorithms produce a higher number of incorrect matches between two photos of Asian and black people relative to white people. Photo: The Washington Post via Getty Images Facial recognition software tests in the United States shows it is biased against ethnic minorities. Algorithms produce a higher number of incorrect matches between two photos of Asian and black people relative to white people. Photo: The Washington Post via Getty Images
Facial recognition software tests in the United States shows it is biased against ethnic minorities. Algorithms produce a higher number of incorrect matches between two photos of Asian and black people relative to white people. Photo: The Washington Post via Getty Images

Facial recognition software biased against Asians and black people, major US government study finds

  • Tests on 189 algorithms from 99 manufacturers, who represent most of the industry, found higher number of incorrect matches for minorities than for white people
  • Use of facial recognition is set to widen at airports worldwide, and travellers may decide it’s worth the trade-off in accuracy if they can save a few minutes

Topic |   Facial recognition
Facial recognition software tests in the United States shows it is biased against ethnic minorities. Algorithms produce a higher number of incorrect matches between two photos of Asian and black people relative to white people. Photo: The Washington Post via Getty Images Facial recognition software tests in the United States shows it is biased against ethnic minorities. Algorithms produce a higher number of incorrect matches between two photos of Asian and black people relative to white people. Photo: The Washington Post via Getty Images
Facial recognition software tests in the United States shows it is biased against ethnic minorities. Algorithms produce a higher number of incorrect matches between two photos of Asian and black people relative to white people. Photo: The Washington Post via Getty Images
READ FULL ARTICLE