I believe that, as far as a complacent company, or an agent in a company, is able to filter the information that people get from the other nodes in their network, the "powers that be" (make that wealth, US goverment, US agencies, whatever fits your bill) can even influence political changes in masses.
That is why the discussion about metadata was so stupid! Politically, metadata IS the ingredient that was missing. One does make political opinions widely available, but metadata allows someone with insight to the network to map influences, make profiles.
And as these two research papers explain, alter their impact in the political process of the mass. It's not the people who are controlled by social networks, but masses surely are:
Exploiting Network Structure in Enhancing Diusion of Complex Contagions: http://www.albany.edu/~ravi/pd...
Effects of Opposition on the Diffusion of Complex Contagions in Social Networks: An Empirical Study: http://link.springer.com/chapt...
Bear in mind that we do not know how edgerank selects information. It could well highlight favourable nodes and muffle problematic ones.
Interestingly, in recent years social movements favourable to western status quo have thrilled in social networks (think maidan, arab "spring", opposition to left leaning governments in South America, now Hong Kong revolts) yet the ones that oppose them have a much larger footprint in the real world than in the virtual world (Chile student revolts, Mexican "I am 132", spanish resistance to shock cuts, that gathered !4million people physically!, Occupy Wall Street). I really wonder if this asymmetry is random or coincidence