We can say that §230 doesn't protect illegal speech but both of those situations seem to hinge on a lack of actual knowledge or intent...but businesses profiting off the results.

It's clearer to me why a website shouldn't be responsible for simple misinformation even algorithmically promoted.