How so? It's meant to be used by researchers distributing the datasets for their own papers. (Notwithstanding the fact that people seem to already be uploading textbooks/lectures, which seems against the stated intent).

Some of the articles seem to be journal articles. Don't get me wrong, I'm not against it, but isn't this partially what Aaron Swartz was indicted for?

Since it isn't illegally acquired and distributed software, no.