While possibly true, we need to attack it from that angle. If you look at a person and go "you are evil!!" you win no one. If you go to someone and say "I'm seeing some behavior in the system you designed. Was this intended? It's harming people, can we fix it?" you recognize them as human and don't assume harmful intent. Then, the both of you can work to solve the problem. If you go at it the other way, you're attacking _them_ and they'll defend by arguing that it's an emergent behavior of the algorithm and there's no evil intended.