Sarah Friend

Year
2018 (ongoing)

Medium / Material / Technic
Video

»Perverse Affordances« examines the political implications of human-computer interaction. Sarah Friend wrote an algorithm that crawled the ten most popular social media platforms worldwide and randomly took screenshots. She then used the prepared dataset of over 10,000 screen captures to train a generative adversarial neural network to produce new screenshots of possible interfaces. The resulting images speculate how the machine learning tools of big social media companies »see« their users all the time.

The title of the work refers to the term »perverse incentive«, which, coming from systems design, means emergent behavior within a system that contradicts the intentions of its designers. »Affordances«, stemming from human-machine interaction studies, means the possibilities for action that an interface offers its user. A perverse affordance might thus be something enabled by an interface but not intended by its designers.

Surveilling thousands of personal pages all over the world, Friend emphasizes that design is never innocent. Contemporary citizens face hundreds of interfaces which, facilitating our activities in data-permeated urban habitats, also frame the set of possibilities for how they might be used by various parties.