You are here

How Algorithms Enforce Personal Roles

For the post-lunch session at AoIR 2015, I'm in a session on algorithms, which begins with a paper by Dylan Wittkower. He suggests that social media experience is influenced by algorithms of display and access on both user and system side.

The most important system algorithm is the newsfeed algorithm that determines what posts from their friends and followers a user engages with. This is based often on social presence - on the recency and frequency of interaction between with specific followers - and thus helps to generate stronger ties between specific users, as well as potentially leading to dominant groupthink. This is because we tend to share things that we have a positive affect towards, and that we think our friends will like – even if we are sharing negative things, we share them in the expectation that our friends will agree with our views.

Further, the topics that produce more interactions are more valued both by newsfeed algorithms and by us as users, which may create a positive feedback loop that reinforces certain sharing behaviours. Normal reward structures for this are shortcircuited by the gamification of such sharing, promoting such activities even more. This leads us to produce and valorise a certain type of profile that we think we ought to have in these social media spaces.

This is a kind of spectacular identity in a Debordian sense: we are packaging our lives as a capital on offer - as "friendtertainment". This is not a false version of ourselves, necessarily, but we end up pigeonholing ourselves for the purposes of each social network, for example by using Facebook for friends and family and Twitter for political discussion.

The second algorithm is user-controlled, as we intentionally create groups. In offline life, physical structures scaffold and enforce specific structures of engagement; in social media, we rebuild such architectures of separation through code by creating groups or circles – a kind of "encircling" specific sets of relationships. This precludes the possibility of weak-tie outsiders becoming engaged in such strong-tie networks, and thus also assigns specific roles to others to perform, as a "standing reserve".