Comment by blacksmith_tb

Maybe, unless people are liking something that AI is providing, without necessarily being aware of it.

Since some people actively try to block AI channels, it's probably quite easy to identify them from user behavior.

It's not hard for us to pick out slop art and writing that's similar to some samples we've seen. We don't need explicit labeling, models shouldn't either.