A couple of years ago I read The Black Stack, and founding compelling his definition of the “inverse panopticon effect”, I toke my phone to add his quote to my records. For that, I used the speech to text feature of the iPhone Notes app, but soon enough I found the task of transferring the text via sound to an interface powered by a machine learning algorithm was almost impossible. Even when the machine heard the words accurately they were changed automatically to what the system understood as more likely to have being said.
These weeks readings provided me another lens to understand the interface: the commons and the infrastructure. The data used to train the interface can be seen as the commons, a shared resource that grows with every user inadvertent contribution and simultaneously affects every user interaction. On the other hand, the infrastructure properties study by Starr can effortlessly be applied to the interface. In my struggle with the iPhone app, the interface bias became visible upon breakdown, upon an interaction that failed to follow the “conventions of practice”.
https://vimeo.com/259922037
Thanks, Liliana! I’m glad to know that some of this week’s central ideas have already been informing your own work 🙂 I’m glad, too, to that the concepts of infrastructure and commons might offer some new lenses for you to rethink some of your past projects. The errors of machine listening certainly do highlight the breakdown of various intertwined infrastructures. Yet I wonder if we can equate data with the commons: are all data presented as a shared resource? We can discuss tomorrow!