An Actor/Structure Perspective on Algorithms

The next speaker at AoIR 2015 is Jakob Linaa Jensen, whose approach is to examine algorithms from an actor/structure perspective. In this, we need to avoid the twin fallacies of techno-optimism as well as techno-pessimism, and move beyond such extreme positions. Algorithms are both good and bad, and a perspective which examines the interplay between actors and structures is useful to shed more light on them.

Actors and structures are ultimately inseparable and inherently intertwined. There are three useful perspectives that may be brought to bear here: Goffman on algorithms and socialisation; Foucault on algorithms and knowledge and biopolitics; and Deleuze and Guattari on algorithms and politics.

Goffman has a dramaturgical perspective on social life that distinguishes front and back stages; mediated sociality is different, and we now interact not only with each other but also with algorithms. We don't know all the rules that regulate our social control, and what we see on social media and how others see us is thus governed by mechanisms that we don't really know ourselves. This adds a new layer of sociality to the system, and Goffman's model collapses.

Foucault suggests that regimes of knowledge are crucial for our operation in society, but algorithms change this game of knowledge: what we know depends on whom we know, and we move into a filter bubble. There is a biopolitical aspect here as well, as we depend increasingly on software agents, not just on actual people.

Deleuze and Guattari's work further points to the fact that we are now experiencing a striated Internet represented by several distinct technological layers into which a range of clashing ideologies from libertarian thought to proprietary possession are built. We are moving into a world of appisation and siloisation now.

As scholars we must investigate and question these developments, as well as to promote transparency and open access to the underlying algorithmic code. We must also distinguish between soft (e.g. Amazon) and hard (e.g. Chinese censorship) algorithmic power, and raise the practical as well as theoretical questions that flow from it.