Sara Goudarzi
In 2015, Krystal Kauffman experienced health problems and could no longer work an on-site job. To pay her bills, she looked on the internet for work-from-home openings and came across MTurk. Short for Mechanical Turk, the Amazon online marketplace allows businesses to hire gig workers all over the world to perform tasks that computers cannot. Kauffman signed up and was soon tagging images and labeling and annotating data for tech companies.
MTurk and platforms like it hire hundreds of thousands of workers worldwide to perform quick tasks, typically for small sums, and they have become an important component of machine learning. In this application of artificial intelligence, machines examine given data to find patterns and use those patterns to learn. But machines don’t always perform as intended, and to aid them, throngs of people take on broken bits of large volumes of work, a phenomenon Amazon founder Jeff Bezos has described as “artificial artificial intelligence (Pontin 2007).”
Kauffman is a research fellow at the Distributed Artificial Intelligence Research Institute (DAIR) and the lead organizer of Turkopticon, a nonprofit organization dedicated to fighting for the rights of gig-workers. She knows artificial artificiality firsthand: “You have all of these smart devices that people think are just magically smart; you have AI that people think magically appeared, and it is people like me [who make them work], and we’re spread out all over the world,” she says.
No comments:
Post a Comment