I’ll ignore the first half of this reply because we won’t agree. Not every choice is a conscious decision in my eyes, but the vast majority are.
As for the second half, believing that bad actors would be weeded out based on the principle of free will is naive. Consider game theory. Two people have something to gain from cooperation, but more to gain from defecting. Meanwhile, the other gains nothing or very little. That simple thought experiment incentivizes bad actions from time to time. You have more to gain by acting selfishly.
Now blow up the experiment. You vs the world and reputation is introduced. Someone with a perfect cooperation rate is flawed. They offer nothing but blind trust and can be taken advantage of. The opposite also displayed. Someone who makes selfish decisions all the time offers nothing but blind distrust. You’re left to choose which people to interact with that are somewhere along the middle of the reputation gradient. Those that are 70% or lower seem unpredictable or untrustworthy so many choose to interact with people on the higher end of the reputation spectrum when available and reflect that in their own decision making. You can’t always choose who to interact with, so eventually you’ll have to interact with a bad actor. You’ll get burned by making a cooperative choice and they will benefit from it. In turn, ensuring that they will survive natural selection.
In game theory cooperation does give advantages.
Both co-op: +1/+1 Both defect: 0/0 Defect/co-op: +3/0
That’s just one interaction. When you expand the experiment, predictability becomes a positive trait and risk is avoided. So by more often choosing cooperation, you become more predictable, avoid the risk of not gaining any points through mutual defection, and more people are likely to interact with you. More interactions=higher potential for points. When you adjust the rules of the game to not define a set number of interactions with each player and you can choose the frequency of interactions with bad reputation players, cooperating is naturally selected for. Conversely, as the pool gets collectively nicer, defection will net more benefits and the pendulum will start to slowly swing the other way.