The origin of the Dissensual Places RtD project goes back to a specific event, which was among the first highlighting the effects of AI-driven social media curation to a wider public. In 2014, well before the current discourses in FAccT, ML interpretability or XAI, an experiment by researchers led by a member of the Core Data Science Team at Facebook generated a short-lived scandal. Kramer and colleagues conducted an experiment on 689,003 (!) users of the Facebook platform regarding a phenomenon well-evidenced in psychology: emotional contagion.
In effect, the study manipulated what Facebook content appeared to users and asked: If you see positive/negative posts, are you more likely to share positive/negative posts yourself? While it is important to argue the ethical dimensions around conducting and publishing this research, my main emphasis here lies on the possibility of conducting it in the first place, and how this gives a particular insight into Facebook’s and similar platforms’ modus operandi. To explicate the latter, the goal in this RtD project was to integrate the phenomenal appearance of the Facebook platform, i.e. the newsfeed interface, with the processes which affect that appearance without becoming part of the experience themselves, i.e. computational processes, into specific artefacts.
To achieve a rendering of the divergent graphic and computational interfaces, I fed recordings of my liking activities on Facebook (recorded using the standard Google DevTools timeline feature) into a custom program I wrote using the Processing IDE (Integrated Development Environment); a popular creative coding tool. I then created 3D prints of the abstract data sculptures. The artefacts of Dissensual Places, as a whole, combine two intersecting ‘worlds’ of sensibility that post-phenomenology needs to account for in order to advance into AI studies: the causal efficacy of technological components, and the experiential interface that is both affected by and affects the former through interlinked feedback loops. Two insights can be derived from this project. First, the resulting abstract sculptures can be interpreted as an argument for the necessity of expanding post-phenomenology’s scope.
Second, Dissensual Places was not only the result of a first-person, practice-led post-phenomenological inquiry. I also exhibited the artefacts in a gallery setting, purposely counterintuitive to the sites of liking I initially investigated and designed with. The presentation of Dissensual Places as an art exhibition provoked highly diverse statements from exhibition visitors. For instance, while there was no declarative explanation that this ‘was’ Facebook, visitors were generally able to derive that the artefacts represented an integration of human and computational activity. The RtD project Dissensual Places, then, proved formative to my PhD research by foregrounding horizonality, i.e. the pre-reflective structures of phenomenological experience, as the key object for further inquiry.