at what point do you think a person would be comfortable with an AI grading them as peer? If you look at things like peer review for example, lots of people complain about the bias and variance per peer. Is there a point at which people would prefer their review done by an AI (which is maybe MORE standard across people and auditable?) Think of the number of things in our society that requires the judgment of peers (at work, court, etc.). Would it be better if it WASN'T our real peers, but something peer-like? The downside obviously is that AI likely can't empathize with a situation that they're grading on a human-level and also doesn't have to deal with the downstream ramifications of their choices (e.g. maybe you don't want to totally annihilate a coworker and actually try to improve them so you can work better together in the future)
1,26K