The University of Pittsburgh’s Mellon Sawyer Seminar, Information Ecosystems, turned its attention to automation and artificial intelligence on Friday, Oct. 25, when the seminar’s postdoctoral fellow Mario Khreiche presented his research related to the future of work in an age of increasing automation.
Khreiche can be described as neither a positivist nor a dystopian. His work lies somewhere in the middle. While he states in his 2019 paper “The Twilight of Automation,” published in Fast Capitalism this fall, “an unchecked project of automation is both ill-conceived and ill-fated” (117) he also takes to task postcapitalist interventions, which he argues “suffers from a certain naïveté, in that its authors undertheorize how emerging technologies unfold as sociotechnical systems, rather than isolated machines” (121). Khreiche is interested in what he calls a “more nuanced question” – something along the lines of trying to figure out how a company like Uber can adapt or change systems to make them less susceptible to technological redlining, for example.
In particular, Khreiche keeps asking what is new about this technological revolution. Work has changed many times in the past: think the Industrial Revolution of the eighteenth and nineteenth centuries or the first computer revolution of the mid-twentieth century. The Luddites of the nineteenth century smashed weaving machines – but not, as is commonly thought and as the term “Luddite” as it is used today indicates, because they were against technology of any kind. Rather, they were concerned about mechanization being used as a way to exploit workers and produce lower-quality goods. This technological revolution, Khreiche argues, will likely be less about replacing and more about displacing, a situation where workers are moved physically and professionally. Indeed, some cultural factors of driving in a specific place would be difficult, if not impossible, to automate: how do you program a self-driving car to know when to take a Pittsburgh left, for example?
What caught my eye in Khreiche’s paper, addressed briefly Friday, was the way he describes Uber’s increased “gamification” of its app, which then encourages drivers to never really sign off and take a break from working. An advertisement to attract drivers claims that Uber allows its drivers to go from “earning” (driving for Uber) to “working” (at the 9-to-5) to “chilling” (as in “Netflix and”) whenever they please. Note that driving for Uber, at least according to the ad, isn’t “working,” it’s “earning,” highlighting the entrepreneurial feel the company hopes to engender.
This constant connection to a job, though, might be harmful. In a viral Buzzfeed News essay, Anne Helen Peterson theorized some of the reasons behind why millennials are the “burnout generation.” The article offers several ideas, but one particularly salient one is the way our smartphones keep us (yes, I am a millennial) working, all the time. One way Peterson points to is that of “branding”: “For many millennials, a social media presence…has also become an integral part of obtaining and maintain a job.” That supposedly nine-to-five job, however, isn’t really, as “the phone is also, and just as essentially, a tether to the ‘real’ workplace. Email and Slack made it so that employees are always accessible, always able to labor, even after they’ve left the physical workplace and the traditional 9-to-5 boundaries of paid labor.”
I got my first Smartphone in 2012, when I started work as a reporter at a daily newspaper owned by a large media corporation. All the reporters had a company iPhone with unlimited data, and we were encouraged to use this phone as our personal phone also. I was also told to get a Twitter account and start tweeting, like, all the time. At the time, I worked Tuesdays through Saturdays, 1 p.m. to 10 p.m., something not all of my sources seemed to understand. As a result, I was emailing sources constantly off the clock, and I, like all my coworkers, were expected to cultivate a “brand” on social media. That work-phone-that-was-also-a-personal-phone quickly became, in Peterson’s words, a “tether” to work. All the reporters, including me, tweeted out photos, noted community goings-on, and provided additional coverage on Twitter, much of it off the clock. (Imagine our chagrin when Twitter introduced its analytics years later, and we all discovered the dismally low audience interaction all those tweets garnered.) In hindsight, my eventual burnout isn’t surprising.
That smartphone forced a connectivity to my job that was never possible before the digital age we currently live in. It also allowed for another ethical concern of automation – invisible labor. In his paper, Khreiche notes that gig economies can “conceal human labor and spin tales of user-entrepreneurialism” (remember that Uber driver who isn’t “working” when he’s driving?). Another gig Khreiche discusses is Amazon’s Mechnical Turk (AMT), which he describes as “a data clearinghouse connecting clients (Requesters) with workers (Taskers or Turkers)” (119). While the cultural narrative claims that computers can do almost anything faster than a human, the opposite is true: “Humans still outperform computers in tasks, such as identifying and classifying objects in images and videos, finding data duplicates, and transcriptions” (120). These are the types of tasks for which AMT connects requesters and taskers. The taskers are, in Khreiche’s words, “kept at the backend of the interface, prefiguring their presence in a twilight economy of automation and residual human employability” (120).
AMT is not, however, a utopia of entrepreneurship and the freelance economy. In a 2017 Wired article, Miranda Katz described workers in India not being paid in a timely manner, and a 2014 effort to organize some turkers called Dynamo. Amazon quickly squashed the effort. Last year, Alana Semuels wrote in The Atlantic about one tasker’s experience on the site, where she sometimes only made $4 or $5 per hour. Like Uber, Amazon’s Mechanical Turk at first may seem like a freelancer’s dream, but the reality may be something different.
The irony, perhaps, is that the name of the service is itself a monumental self-own. The “original” Mechanical Turk was a phony automaton built in the 1770s. Its inventor, Wolfgang von Kempelen, claimed the Mechanical Turk could play chess against a human. The automaton was a hit in the Vienna court of the Archduchess Maria Theresa, but it wasn’t all it seemed. In order for the machine to work, a human had to sit in a cabinet underneath the chess board and direct play. This bit of invisible labor delighted and sometimes troubled onlookers. Amazon’s twenty-first century version also relies on invisible labor, with a hope that nobody bothers to peek underneath the cabinet and check on the cramped, sweating person behind the machine.
Briana Wipf is a second-year PhD student at the University of Pittsburgh. She listens to the Beatles while composing these blog posts. Follow her on Twitter @Briana_Wipf.