Good catch. Notably, this is pre-replication crisis, and the evidence is kinda underwhelming anyway for the claims that are being made. The theory that increasing the salience of personal power makes experimental subjects less empathetic could've been tested more directly than that. And as for a practical application to things like fairness and pro-sociality, what we know about e.g. the well-established literature on "Dictator" vs. "Ultimatum" games strongly suggests that people in a position of power are only marginally less pro-social/"fair" to others, if that.