I'm assuming this is a big improvement in their duplicate content filtering. But I'm also thinking that maybe this means that those non-indexed copies generate little to no SEO value?
Duplicate content filtering and non-indexing are two different things happening for different reasons. Linking one to the other as related may not be helpful to understanding what is going on with article marketing. Here's why:
Duplicate content filtering is simply the process of identifying which document is the most authoritative and original source. That has little or anything to do with whether the content is indexed and even less to do with whether the links will pass full, partial or no PageRank.
As far as non-indexed copies, you'd be surprised at what is actually indexed, by copying a portion of the content, wrapping it in quotes and doing a verbatim search. If the page is truly not-indexed, what that really means is that it is in a state usually associated with being banned. So if you say something is not indexed, that's a very severe state, something that is usually seen in web pages that are actively and aggressively being excluded. Being excluded and being filtered out are two different things. Just because it doesn't show up in the regular serps doesn't mean it isn't indexed. It just means it doesn't have enough PageRank/authority to be useful in the SERPs, usually referred to by some as being filtered out.
Now that we hopefully understand that filtering web pages from the SERPs is not the same as non-indexed, we can better understand that just because something is deemed a duplicate does not mean that the links have no PageRank or authority of their own to pass along.
I think the links are not passing authority, but not because the pages are duplicate content. I say that based upon SERP observation. I will explain this in another post. But if anyone wishes to add feel free.