I don’t think that’s a big problem, “agree” would be interpreted as “mostly agree”, and if many people disagree with only one of the points, there will be a highly upvoted/agreevoted reply.
The halo effect causes people to judge things as good or bad overall, estimating all aspects as better or worse than they actually are. Making people choose whether a post is good puts the points that it contains into a mental category, which makes people more prone to this bias. An AI could decide whether it mostly agrees without any effects on its juddment of individual points, but humans are quite a bit more fallible.
I don’t think that’s a big problem, “agree” would be interpreted as “mostly agree”, and if many people disagree with only one of the points, there will be a highly upvoted/agreevoted reply.
The halo effect causes people to judge things as good or bad overall, estimating all aspects as better or worse than they actually are. Making people choose whether a post is good puts the points that it contains into a mental category, which makes people more prone to this bias. An AI could decide whether it mostly agrees without any effects on its juddment of individual points, but humans are quite a bit more fallible.