> I can imagine AI summarizes being problematic for a class of people that do not cross check if an article is of value to them.

I feel like that’s “the majority of people” or at least “a large enough group for it to be a societal problem”.