Regenerating Algorithms

Abstract

This panel will bring together a number of short presentations that address how algorithms—originally designed for purposes ranging from property valuation to pregnancy detection—might be regenerated to support humanistic values such as critical reflection and social justice. These presentations will illustrate a range of ways in which we can remake algorithms: rethinking the choices and authority of their training data; recalibrating their functionality for specific contexts; and realigning their outputs for new audiences. This regenerative design exercise leverages a long history of potent methods and concepts from STS literature for understanding the situatedness of information technologies, while at the same time turning them toward practical ends.

We take as our starting point the recognition that all algorithms are local and value-laden: created by people and their dutiful machines, at a time, in a place, with specific values and politics, for audiences that are conditioned to interact with them. Each panelist will respond to an existing algorithm used for one of the following: property value estimation, networked medicine, pregnancy detection, natural language processing, digital navigation, and even just "sorting." Panelists will develop their own means of representing the original algorithms, as well as regenerated versions and the potential implications of their proposed changes. Finally, we will leave considerable time at the end of the panel for a moderated discussion that connects the algorithms and panelists’ strategies for representation and regeneration.

References:

JafariNaimi, Nassim, Lisa Nathan, and Ian Hargraves. "Values as hypotheses: design, inquiry, and the service of values." Design issues 31, no. 4 (2015): 91-104.

Loukissas, Yanni A. All Data Are Local: Thinking Critically in a Data-Driven Society. Cambridge: MIT Press, 2019.

Ratto, Matt, Sara Ann Wylie, and Kirk Jalbert. "Introduction to the special forum on critical making as research program." The information society 30, no. 2 (2014): 85-95.

Suchman, Lucy. "Frankenstein’s Problem." In Working Conference on Information Systems and Organizations, pp. 13-18. Springer, Cham, 2018.

Chair

• Nassim Parvin, Georgia Institute of Technology

Individual Presentations

• Re-Valuing Homes: Countering the Culture of Real Estate on Zillow - Yanni Alexander Loukissas, Georgia Institute of Technology

• Plans and Situated (Algorithmic) Actions: Networked Medical Devices as Sites of Feminist Technoscience or Encountering the More Than Human in the Middle of the Night - Laura Forlano, Illinois Institute of Technology

• Dissident Pregnancy Detection Models for Reproductive Justice - Catherine D'Ignazio, Massachusetts Institute of Technology

• "Earth Measurer": Visualizing a Recurrent Vanilla Neural Network, Figuring Abundance - Matthew Battles Battles, Harvard's Berkman-Klein Center for Internet and Society

• Navigation through an Environmental Lens - Jude Mwenda Ntabathia, Georgia Tech

• Visceralizing Algorithms - Charles Luke Alan Stark

Session Organizers

• Yanni Alexander Loukissas, Georgia Institute of Technology

• Nassim Parvin, Georgia Institute of Technology