Prof. Irani's talk will focus on her research with Amazon's Mechanical Turk platform and the broader questions it raises about justice.
Lilly Irani has contributed to the design and maintenance of Turkopticon, a tool and web service that enables Amazon Mechanical Turk (AMT) workers to share job information with one another and hold employers accountable for fair treatment. Turkopticon is a browser extension that inserts reviews of employers into AMT’s task marketplace interface. Thousands of workers a month use the system.
AMT is a system that allows programmers to solicit large numbers of workers to do small bits of computational labor, such as transcription, image labeling, or survey participation. These bits of labor can be integrated directly into computational systems, layering human computation into existing and emerging information technologies. Human computation represents a growing area of both academic research and industrial development.
Prof. Irani's work examines what happens when people are organized as smooth computational infrastructure, integrated directly into algorithms. As a source of ethical insight into AMT, her research group began by surveying workers, asking them to draft what they would like as a worker’s bill of rights. Though workers’ responses diverged wildly, many agreed on one problem: they had little recourse when employers withheld pay. Her group responded to these findings by designing Turkopticon.
Through Prof. Irani's work on Turkopticon, she has explored how communications technologies support or hinder digital solidarities. Through the design and maintenance of the system, she has also explored the politics of mediated and massively distributed participation in design.