Difference and Dependence among Digital Workers: The Case of Amazon Mechanical Turk

A G A I N S T the D A Y 1 Lilly Irani Difference and Dependence among Digital Workers: The Case of Amazon Mechanical Turk I n 2006, Jeff Bezos launched a new form of computer technology. Amazon.com, Inc. had used the technology as a form of “artificial artificial intelligence”—data processing that could classify images, sounds, and texts automatically while still seizing on cultural nuances like humor, sexual- ity, and linguistic dialects. The service was part of Amazon Web Services, marketed alongside S3 and EC2 1 —just-in-time server space and computa- tional cycles available to programmers through routine acts of coding. Bezos explained the new technology—the artificial artificial intelligence—as “humans-as-a-service.” That service was Amazon Mechanical Turk (AMT). The secret of AMT was not a feat of computer engineering, statistics, or algorithms. In fact, AMT was born out of the failures of artificial intelli- gence to meet the needs of internet companies seeking to expand the domain of the data they could store, classify, and serve up online. Rather, AMT offered a virtual marketplace where workers with computers and internet connections all over the world could flexibly complete data-processing tasks around the clock. Employers seeking quick-turnaround data processing no longer had to hire more employees or even contract with an outsourcing firm; they would not even have to meet their employees, either online or face-to- face. They could simply place their data-processing tasks online, set a price for each task, and design algorithms to receive, validate, and integrate work- ers’ processed data into computer systems. The system allowed for a kind of massively mediated microlabor—large volumes of small, independent tasks distributed to large groups of workers. The South Atlantic Quarterly 114:1, Winter 2015 doi 10.1215/00382876-2831665 © 2014 Duke University Press SAQ114_1_16Irani_1pp.indd 225 9/19/14 2:08 AM