There are big problems with facial recognition programs. Researchers have found 35% of women, and especially women of color, were misidentified, while white men were identified 99% of the time.
Click above to listen to part 2 of Skeptic Check: Data Bias, featuring Kade Crockford, Director of the Technology for Liberty Program at the ACLU of Massachusetts, discussing how facial recognition systems fail to identify or misclassify images of faces, overwhelmingly so with faces of people with dark skin.
Skeptic Check: Data Bias:
Part 1: Caroline Criado-Perez / Invisible Women
Part 2: Kade Crockford / Facial Recognition
Part 3: Amy Webb / Proprietary A.I.
