Abstract
Abstract
This article investigates the rise of Amazon Mechanical Turk (MTurk), Amazon Web Services, Inc.’s crowdsourcing labor platform, in social science research since 2005. A new “digital sweatshop,” the platform hired online workers to do precarious, extremely low-wage tasks to support artificial intelligence (AI) and survey research, while effectively stripping workers of all protections except those they built for themselves. Bringing together labor history and the history of science through an investigation of MTurk, this article intervenes in the historiography bidirectionally. Interpreting research participation as work, it argues, first, that the history of knowledge production is a labor history. To understand the ensuing conflict between workers and researchers on the MTurk platform, one must understand its labor context. Their struggle lay at the intersection between social science's notion of ideal research subjects and the concerns, interests, and vulnerabilities of crowdsourced participants as a class of exploited and unprotected workers. This article asks, second, how the labor conditions of research subjects impacted the knowledge produced from them. As in other industries, dialectics of labor exploitation shaped (and spoiled) the knowledge products that digital piecework yielded. The “labor” being deskilled, in this case, was being human.