> Men have long dominated fields such as finance and science, for example, so the technology would conclude from the data that an investor or engineer is “he” or “him.”

Err, so the AI correctly predicts that statistically you probably intend "him", but we limit the utility of the tool because that would be discriminatory? I know it will get it wrong sometimes, and you can say it re-inforces stereotypes, but if it will get it right most of the time based on strict statistical inference, seems like it could at least be configured. It seems to be a case where the AI is a bit too accurate... I don't disagree with their decision, I think given the circumstances it's actually a brilliantly safe move and keeps out of the fray as much as possible.

> “The only reliable technique we have is to be conservative,”

I know this is just whimsical, and a horrible logical equivocation on my part, but that's kind of funny that a large tech company decided being conservative might actually be useful in some cases to help protect against liberal outrage...