Algorithmic Worldmaking

My current project is titled Algorithmic Worldmaking. The project focuses on the role of algorithms in technological systems. I am interested in understanding how algorithms become facilitators and subjects of public contestation and controversy. I contend that algorithmic power comes from shaping the flow of rhetorical exchange. Algorithmic agency comes from directing the flow of attention and materials in networked media. Digging into ancient Greek theory and philosophy, I argue that the term kosmos best encapsulates how socio-material order flows in networked systems. Kosmos, which can be translated as "order," was a negotiation among humans, technologies, and environments. Algorithmic kosmos, I argue, manifests in a few primary patterns: navigation, exploration, maintenance, and monetization.

  1. "Navigating Networked Worlds": This chapter unpacks how algorithms reshape networked spaces by helping users navigate from point to point. I begin with Google search, which I argue is a process of navigation, one that can be disrupted through human intervention. I study the Google bombing of Rick Santorum, a campaign designed to defame the former Senator by offering an explicit alternative definition of Santorum. I also analyze the alt-right's use of Google bombing to spread white nationalism and I trace the manipulation of mapping systems to spread anti-Semitism.
  2. "Exploring the Agora": This chapter lays out processes of exploration, which are designed to help users find new things rather than merely moving from point a to point b. I study recommendation and targeting algorithms, articulating a theory of "personalized publics" and considering the algorithmization of identity. I trace Facebook's deployment of "ethnic affinity advertisement," which allowed the site to push different advertisements for Straight Outta Compton, and which came under fire for allowing advertisers to illegally restrict minority users from viewing housing ads.
  3. "Maintaining Communities": This chapter examines how algorithmic systems maintain the structure of communities, often in ways that reify the power structures of the "offline" world. The chapter primarily considers moderation algorithms and procedures. I first assess Facebook's algorithmic moderation system and its tendency to protect white users while restricting Black voices on the platform. I then analyze the "counter-information" efforts of info panels on YouTube before diving into the destruction of Tumblr as the platform leveraged algorithmic moderation to remove adult content.
  4. "Monetizing the System": This chapter articulates the harms of monetization algorithms, which underpin most of the other types of algorithmic kosmos studied throughout the book. Monetization algorithms are designed explicitly to extract value, including copyright matching algorithms (Content ID on YouTube and Audible Magic elsewhere) and gig work algorithms, specifically DoorDash's, which were the subjects of a campaign (#DeclineNow) of manipulation designed to create better wages for workers.

In each case, algorithms create a world that reflects the human biases of the "offline" world. Algorithms are not inherently objective, unbiased, fair, or just. Rather, we must recognize their tendencies to harm minoritized users, reify privilege, and prioritize profit. I offer in the conclusion a few strategies for creating better algorithms, focusing on a simple notion: to improve our algorithms, we must improve ourselves.