Facial Recognition Misidentifies California Lawmakers as Criminals

By Tenta August 18, 2019

The American Civil Liberties Union (ACLU) of Northern California recently ran a test of Rekognition, Amazon's facial recognition technology, and found that it misidentified 26 California lawmakers as criminals. The misidentified lawmakers were a portion of the 120 lawmakers that the ACLU scanned against images of 25,000 criminals. San Francisco Assemblyman Phil Ting was one of the ones identified as a criminal, despite never having been arrested.


This is a companion discussion topic for the original entry at https://tenta.com/blog/post/2019/08/facial-recognition-misidentifies-lawmakers-as-criminals