Not that traditional journals are exempt. There are simply too many journals, and good peer-reviewers are overworked, meaning that a lot of stuff gets relegated to potentially unqualified reviewers. And the pressure to publish N papers per year has led to a "let's throw it at the wall and see if it sticks" attitude to journal article submission. Science suffers from all of this noise, and we have only ourselves to blame for falling into the bibliometric trap set for us by the bean-counters.
Which makes me wonder. Should I review articles from groups I know and trust to be careful and honest, and nit pick their papers to make sure they are rock solid and bullet-proof? Or should I spend time instead reviewing for lesser journals and novice authors, where the reviewer is implicitly being asked to teach the authors how to write a paper, to make sure that crap doesn't inadvertently make it into the system?