If it's a sentient ASI (instead of a glorified program that can give you the right answers but doesn't actually know what these answers mean) it all depends on how altruistic it programned itself (an ASI would be able to change how its own mind works) to be.

If it's a lifeless program, it depends on how much the people that have access or own ASI care about those that don't have access to it. But honestly I don't think they would be any different from the rich assholes we have today.

case 1. maybe

case 2. no, humans are fking evil