>We are now finding algorithms that have implicit racial biases because the only people who worked on them were Asian and Caucasian How do you know that these

>We are now finding algorithms that have implicit racial biases because the only people who worked on them were Asian and Caucasian How do you know that these biases aren't just representative of actual biases in reality? More importantly, given the relatively small proportion of AI applications that operate on human data, indication of such a bias is a poor argument for the value of diversity across the industry.