As technology becomes more sophisticated, they will have more potential to check for spun content in my opinion.
Imagine it like code breaking, article spinning software *encodes* the article, using a *template* (or set of synonyms), Google has it's own set of synonyms, and can decode this into possible variants, then when it finds a similar variant that is a certain % close enough, it runs further checks for footprints, such as:
Do they link to the same site/s?
So then say they find 100 articles, all with the same *encoding*, all linking out to the same domain.
The odds of that naturally are 0.0000001%, therefore these sites are affiliated, or compromised, so they either discount the links, penalise the linking domain, or penalise the article domain.
On mass, it all has to be scalable, and there has to be thresholds.
More advanced footprints, using the *average sites* as a middle ground, is their way forward.
So small scale, manually created networks will never be caught, as their is no true way to find a footprint for something that is truly original.
However, if you cut corners, you run the risk, and whereas spun content might not be trackable now, in the future it will be.
