Avant-Garde Procedural Art Week 2: Meet your OCEAN, an instruction-based participatory performance where you perform someone else's digital footprint

One of the most widely discussed topics lately in tech circles is the role that data science and AI played in the election, specifically the part played by Cambridge Analytica, the political marketing company behind 2 biggest upset campaigns of the past year – the Brexit campaign and Trump campaign. By correlating social media behavior (likes, posts, link clicks, etc.) to psychological traits, Cambridge Analytica enabled to produce ads for different groups fitting different psychological profiles (e.g. the highly anxious, or the mightily patriotic), and to accurately identify voters who could be swayed to vote for their candidate, or swayed to not turn out to vote for the opposing one.

OCEAN, the psychological model employed by Cambridge Analytica

Even though claims that Cambridge Analytica was solely responsible for the election results are exaggerated, it is undeniable that huge danger lies ahead as corporations become able to model our behaviors with ever-increasing precision, achieving a level of insight in how we behave that could surpass our own self-knowledge:

Seventy “likes” were enough to outdo what a person’s friends knew, 150 what their parents knew, and 300 “likes” what their partner knew. More “likes” could even surpass what a person thought they knew about themselves.

Antipersona (2016), an app I made that lets you enter someone else's Twitter account

I’m very interested in those developments because over the past 2 years or so, I’ve been making works about identity and technology, specifically looking at how technology will enable us to create new identities and enter the identities of others. Much like Cambridge Analytica, I see behavioral data derived from social media as a lens (however imperfect) to peer into someone’s subjectivity. I depart from them in my desire to use that data to bridge empathy gaps and allow people to better understand each other.

"The Essential Guide to Performing Michael Mandiberg" by Michael Mandiberg (2002) consists of a set of guidelines for anyone who is interested in inhabiting the identity of Michael Mandiberg.

I think that instruction-based art could be employed to critically study this kind of mass psychological profiling performed by data analytics operations. If to understand is to simulate, and to simulate is to execute a fixed set of instructions, then would it be possible to “perform” a digital simulation of someone else? It seems to me that physically “embodying” a digital simulation would be a very effective way to examine it, identity its biases, see the ways in which it is eerily accurate, the ways in which it falls short.

In the episode "Be Right Back" of the sci-fi series Black Mirror (2013), the protagonist discovers that the simulation of her dead boyfriend, which was constructed from his social media footprint, does not have his sexual preferences, because he never posted about them online.

To test that hypothesis, I will create a participatory piece where you get to have a conversation with your digital footprint. After each participant consensually provides their Facebook profile data, I will employ IBM’s Personality Insights API to derive a psychological description of them based on their Facebook posts. Additionally, I will collect their likes and employment/education/location timeline. Based on those pieces of information, I will generate a dialogue between pairs of participants, each of which will be performing each other digital’s footprint. The goal here is twofold:

1. Getting to interact with the digital simulation of yourself (=> becoming more aware of what companies and governments understand about you)

2. Putting yourself in the shoes of someone else’s modeled identity (=> escaping the comfort zone of your identity)

The IBM Personality Insights API includes (and goes beyond) OCEAN analysis.