Algorithms, Public Culture, and Platform Responsibility

Session Submission Type: Panel

Abstract

This session is occasioned by Vilém Flusser’s observations about the relationship of culture, information technology, and critical practice. Writing in 1985, he speculated that computers would soon absolve human beings of much of the responsibility for selecting, evaluating, curating, and transmitting culture. Sophisticated software applications, combined with faster, more powerful processors, would help inaugurate a new class of “automatic critics and censors” tasked with determining which aspects of human expression most deserved to rise above the din. Flusser was cautiously optimistic about this shift. He noted that computers would take much of the immediate work of processing culture out of people’s hands, thereby allowing us to engage in a more broad-ranging conversation about the “structures of metajudgment,” i.e. the critical frameworks, that would guide computational-cultural decision-making.

The world of “automatic critics” Flusser envisioned has materialized, though maybe not as he had expected. Today, computer algorithms play a decisive role in shaping many aspects of public culture. They surface relevant aspects of news and information; facilitate friendship, romance, professional networking, and other types of interpersonal encounter; tailor collections of books, videos, and music to specific communities; dissect consumer behavior for better-targeted advertising; and more. Yet, these developments have not been met with equally rich opportunities to architect and debate the critical frameworks by means of which computer systems make decisions about cultural life, cultural value, and cultural goods.

Over the last 20 years, online platforms including Google, Facebook, LinkedIn, Twitter, Amazon, Netflix, Reddit, and Pandora have assumed the bulk of this responsibility, their algorithms defining values and procedures for parsing and judging billions of instances of culture each day. Yet, their algorithms are also proprietary. The inner-workings of these systems are shrouded by non-disclosure agreements, trade secret laws, corporate non-compete clauses, and other legal instruments. Conversations about metajudgment thus are happening less in public than behind closed-doors, a situation that begs the question of platform responsibility: what happens when the critical frameworks for assessing public culture become automatic, proprietary, and opaque?

Gillespie opens the panel with an overview of the idea of “platform responsibility,” a key conceptual touchstone for the panel. Striphas lays the other conceptual touchstone for the panel in providing a critical-etymological overview of the keyword, “algorithm.” Marwick then takes an up-close look at how algorithmically-generated metrics promote inequalities on popular social networking sites. Sandvig concludes the session with a critical look at machine learning, or the “training” of algorithms to satisfy the tastes of consumers of cultural goods.

Sub Unit

• Philosophy, Theory and Critique

Chair

• Ted Striphas, Indiana U

Individual Presentations

• Platform Responsibility, and the Curation of Public Discourse - Tarleton L. Gillespie, Cornell U

• What is an Algorithm? - Ted Striphas, Indiana U

• Your Fave Could Never: The Political Economy of Algorithmic Popularity - Alice Marwick, Fordham U

• You Are a Political Junkie and Felon Who Loves the Sound of Blenders:
 Machine Learning Taste Publics - Christian Sandvig, U of Michigan, Ann Arbor