Well, in this study, they explicitly had equality - all four groups were as likely to succeed. So any difference in hiring was actual bias (or statistical noise).

The point is about the definition, not about how it relates to the specific circumstances.