AI programs (real or pretend) often implement racial bias. California is considering a law that attempts to stop this -- by identity politics!

Identity politics is not a real solution to any problem. What's more, the problem of bias in algorithms is subtle. It is a mistake to assume that people know how to avoid it or even detect it as a consequence of their demographic background.

What California (and all polities) ought to do is require all software used to make decisions about individual people to be free software and its source code published, That way the public will be able to study whether its behavior is biased.