Why :i:like green

2020 |

Why an algorithm categorised me by liking ‘Sportswear,’ ‘Prefectures of Japan,’ ‘The American Mathematical Society,’ ‘Walter,’ ‘Vegan nutrition,’ ‘Freight transport,’ ‘Green,’ – and 114 other things…

Research and project description

Why :i:like green

In our physical world, we understand identity. We know when to behave like a friend or lover, professional or student, a parent or child etc. Intuitively we change roles all the time. In a digital domain, we can’t secure our identity that same way: while we carefully curate our social media– and dating profiles and think we choose what to share, conversely, we are captured in predictive and risk profiles. Summarized in data and based on algorithms, preserved by companies and institutions. This is called ‘profiling’.

Where is the data coming from, and how is this information converted into profiles? For example, Janssen found out that a Facebook algorithm categorized her into 114 interests, such as musea, sportswear, technology, prefectures of Japan, mathematics, and the colour green. “Why does this system think that I’m interested in the colour green?” she wondered. This triggered many questions like; does she favour green over other colours, and if so, what might be indicators for this algorithm to address her digital identity?

In this project, Janssen searched for answers and mapped the infrastructure of how Facebook gets access to her data beyond the platform itself (like all event pages you have ever visited). This is happening in four ways: Over the years, Facebook bought multiple companies, like Instagram and Whatsapp, to ensure access to more data flows. Moreover, they have admission to all cookies that are active on your devices. Also, the social media platform provides Third-Party Authentication, allowing you to use your Facebook account to log in to other platforms, such as Spotify or Tinder, which serves them the data from these platforms too. And finally, they place pixels (a technology similar to a cookie) on websites, e.g. supermarkets, warehouses, political websites, to track user behaviour into the Facebook profiling system.

Janssen collected all the data points, designed a system (with colour and typography) to clarify the different data sources and worked out a technique to print these, with a robot, on ping pong balls. The installation consists of almost 4.000 ping pong balls – just a tiny part of all the information Facebook has access to. Why :i:like Green, volumizes the intimacy and massiveness of private information. The installation invites the audience to play a Bingo game with Janssens’ data to find the data that might imply why she likes green.
Therefore, this project questions the possible complications concerning our integrity and autonomy due to profiling. Will this have long-term consequences for the choices we make in life? Because we don’t know what data led to what conclusion. Can we, humans, understand how algorithmic decisions come about? And what impact they might have over individuals and, on top of that, on our society.