Networks, Networks, and More Networks: But Are We Benefitting?
Facebook is in the midst of unveiling its latest set of changes, including a partnership with web music service Spotify and a new feature showing real-time updates of user’s activities. While it’s the imminent profile revamp that will probably incur the strongest reaction, these new connection and reporting functions bear thinking about for what they represent: Facebook branching out to other Internet sites. Facebook has always promoted sharing—oversharing, and with potential privacy violations, according to some—but increasingly, it’s internetwork and not intranetwork. Facebook wants us to link our FB accounts to the other web sites and services we use. It does this explicitly through connections like the one to Spotify, but also much more subtly through those ubiquitously placed “Like” buttons on non-Facebook sites. When we “Like” or “Recommend” something, we’re sharing it with our Facebook network; when we don’t actively click on a button, it must be noted that any site with one of those buttons is still collecting information about our visit and sending it back to Facebook. In essence, we’re adding another network to our existing Facebook network of users (nodes) and friendships (edges).
To be clear, this is very different from creating an account on a different social networking site such as Google+. Then, we are establishing a network separate from our Facebook network, even if it’s modeling the same real life relationships—starting a new network graph in which the nodes we’re linked to and the way we’re linked to them is not necessarily the same. Overlaying a new network on a different site to our Facebook network loses many potential benefits because it is redundant, drawing on the existing structure. The limitations are comparable to the failings of the mathematical explanation of the small-world phenomenon: expanding one link further doesn’t help us reach anywhere if we are already share all that node’s connections in common. This is generally what the user faces when linking accounts, taking as an example Spotify and Facebook. For users, the key part of the social network is other users as nodes and friendships as links. For the companies however, we can generalize the network, seeing every interest as a node and every expressed preference as an edge. Partnering up means that Spotify gets access to your friends and Facebook gets access to your music listening habits–clear advantages that can generate a set of very valuable network data, exactly the kind of information advertisers want to know.
Perhaps even more alarming is all the networks we are enrolled in by default policy, but that we don’t know about. The automatic data collection such as that done by “like” buttons is generating a network that we are part of—except that we don’t have any information about the network, nor can we interact through it. A recent uproar arose over a revision in OnStar policy, which would have let the company collect and possibly sell vehicle data such as location, speed, etc. from customers who had cancelled their OnStar subscription. OnStar has since agreed to retract this change, though only after Senator Schumer filed for an investigation of the practice by a regulatory commission. So who benefits from this increase in networks and network information, if it’s not the users (as the drivers in the OnStar example clearly don’t)? It’s Facebook and the other data mining companies that get to use or sell all this network knowledge. Maybe Facebook isn’t the medium through which the network functions—maybe it’s more accurate to say Facebook itself is a node, linked to every user. If so, Facebook is a hub in the Internet network: one that is connecting us to more networks every day, sometimes without our knowledge, and gaining power for it.
http://bits.blogs.nytimes.com/2011/09/27/as-like-buttons-spread-so-do-facebooks-tentacles/
http://www.cnn.com/2011/09/23/tech/social-media/facebook-real-time/index.html
http://www.wired.com/threatlevel/2011/09/senator-onstar-brazen-privacy-invasion/