By Hyeonjeong Joo

Associative keyword algorithms — such as Google’s suggested search or related image results — provide convenience and efficiency, but ultimately serve to encapsulate the user in recursive isolation. The more such algorithms inform their results through pattern recognition and profiling, the harder it becomes for the user to find novelty, diversity or an escape from enforced predictability.

In Orange is the New Black, Hyeonjeong Joo collects the metadata that shapes her personalized search results as a means of interrogating which biases are introduced by Google’s algorithms. Scanning through her cloud-based image archive, Orange is the New Black retraces the steps of algorithmic machine vision to speculate on which visual clues may have informed the biases in her search results.

The research for this project is developed from two related sets of digital data: the associated keywords produced by Google’s search engine and the artist’s personal image archive stored on Google Drive.

Does associated search or automated phrase-prediction through algorithms save us time? Or does it waste time? The associated keyword algorithms provided for convenience seem to suggest a new direction, but eventually seem to serve only quantities of similar results.

Do such algorithms risk trapping us within the boundaries of a shared bias or in fragmented thoughts? What is the impact on diversity?

An algorithmic system built upon our personal data creates a warped digital reflection. Through this digital persona made from our data, this research attempts to show why we should be more wary of our digital consumption by materializing the dangers of prediction.

[1] Boris Groys (2013) Art Workers: Between Utopia and the Archive. Available from https://www.e-flux.com/journal/45/60134/art-workers-between-utopia-and-the-archive/ (accessed 26 May 2021).

[2] Kim Hyunja (2020) The algorithm makes you a frog in a well. Available from http://www.ohmynews.com/NWS_Web/View/at_pg.aspx?CNTN_CD=A0002694154 (accessed 26 May 2021).

[3] Swathi Meenakshi Sadagopan (2019) Feedback loops and echo chambers: How algorithms amplify viewpoints. Available from https://theconversation.com/feedback-loops-and-echo-chambers-how-algorithms-amplify-viewpoints-107935 (accessed 26 May 2021).

[4] C. Thi Nguyen (2019) Building Trust As A Solution To Breaking The Infinite Loop Of An Echo Chamber. Available from http://www.milwaukeeindependent.com/syndicated/building-trust-as-a-solution-to-breaking-the-infinite-loop-of-an-echo-chamber/ (accessed 26 May 2021).

Search history using categorization in google (keywords)

First outcomes

During the research, various visual strategies were explored to articulate the process of machine learning and vision. The notion of an inescapable sinkhole is used as a signal to warn of the risk inherent to predictive algorithms in erasing objective thinking or perspective.

Sinkholes may emerge as a result of natural causes (such as continental drift) or through excessive urban development.

A looping effect is produced through reflection using a mirror: the images reflected in a mirror seem to show an infinite recursive loop but ultimately just repeats itself in a finite space.

Installation sketches

Installation tryouts

The image that is sucked into the dark sinkhole is made out of four parts in the video, and mirrors are attached to two side of the construction to show a repetitive image cycle.