Much attention has rightfully been given about how to the AI industry might transmit existing negative biases into the myriad of artificially intelligent systems that are now being built.

Bias typically arrives along with real world data; as long as our society exhibits negative biases, we’ll see these reflected in the data we collect. Eliminate such bias from the dataset is unrealistic.

There is another option. Today we can build AI that reinforces positive attributes a reality through programmatic product design choices, it isn't easy, but it is possible. This means introducing ethics in programming. Should be this desirable? Who will decide?

Food for thought.

See the article by  Matthew Hutson "Even artificial intelligence can acquire biases against race and gender"