The Ghost of the Mechanical Turk
Digital microwork in the Middle East exploits occupation, war, and neoliberalism to extract the cheapest labor possible.
“Be yOuR OwN BosS.” Two twenty-something Palestinians are pictured leaning over an iPad and laughing: a ColourSplash™ filter makes their eyes and frayed festival wristbands glow a radioactive green. Their suits are whatever they put on that morning and their offices are wherever they turned their screens on. On a Facebook group for online freelancing work in Gaza and the West Bank, these mantras come up again and again.
The idea is that regardless of your circumstances, anyone can live the millennial dream and “work where they want when they want” thanks to the internet. In a place like Palestine, where unemployment reaches 30 percent (the highest in the world by some measures) and movement is violently restricted by a series of checkpoints, borders, and military zones, work in the “placeless” digital realm can be sold as a way of overcoming these obstacles.
The World Bank’s “m2work” project in cooperation with Nokia took precisely this approach. It was just one of a series of initiatives in the past few years led by governments or private-sector actors that have identified the Arab world as a region in which digital microwork has “vast potential” as a means of alleviating poverty. But this rhetoric of flexibility and entrepreneurship conceals some ugly realities.
There are impressive grassroots projects such as Gaza Sky Geeks working on digital labor initiatives with more emancipatory potential. But when Western bodies like the World Bank, the Rockefeller Foundation, or multinational NGOs tout gig-economy platforms as a silver bullet to the problems of refugees and victims of occupation, their motivations must be examined. “Liberation” from a techno-developmentalist’s view looks more like exploitation from the worker’s point of view.
Microwork, key to the neoliberal development schemes targeted at the Middle East, is best exemplified by Amazon’s “Mechanical Turk” labor platform. As David Golumbia wrote in Jacobin in 2015,
[Mechanical Turk’s] name is . . . a revealing reference to the chess-playing automatons of eighteenth-century Europe, a parlor trick concealing small human beings who actually did the work purportedly done by machines. As historian Ahyan Aytes notes, these automatons were dressed in “Oriental” garb in part because everyone to the East of Europe was understood to be “docile” and “soulless.” MTurk allows employers to design tasks that require large amounts of data entry and analysis that, for whatever reason, currently remain more efficiently or more accurately done by human beings than by computers.
Microwork portals harness virtual crowds to organize playlists of music tag videos and images, write, and translate or transcribe short texts; and, in doing so, to train artificial intelligence software.
Microwork is just one part of a broader spectrum of digital labor that ranges from on-demand services like Uber to the extraction of profitable data from our casual Facebook, Twitter, and instagram updates. What sociologist Antonio Casilli represents as “a continuum of unpaid, micropaid, and poorly paid ‘taskified’ human activities” means that work can no longer be easily distinguished from leisure time. Because of this, it’s hard to talk about exploitation, a word usually associated with industrial labor’s sweatshop conditions, in relation to digital labor. Its also hard to reconcile people not always feeling like these activities are work with their objective creation of a great deal of economic value.
But microwork isn’t innocuous just because it might be completed with the same keystrokes used to chat with a Facebook friend or download a movie. Studies by Human Computer Interaction (HCI) researchers, most notably Lily Irani and Six Silberman and the Oxford Internet Institute, have analyzed these platforms and conducted interviews with users to show how paid digital labor platforms seriously threaten basic workers protections. While the median hourly rate on Amazon Mechanical Turk is $1.38, tasks are completed for as little as $0.01. Wage rates plummet as people around to world bid to complete tasks for cheaper and cheaper prices. Opportunities for promotion or skill-development are near non-existent as workers are unable to access information about core business processes.
And digital workers’ bodies are no less regimented to the demands of capital than their industrial counterparts. Refugees in Lebanon’s Shatila Camp, who engage in a range of digital labor activities from freelance programming to online games with virtual currencies, are forced to adjust their sleeping patterns to the demands of (western) capital. Teenagers keep their phone buzzers on loud while they sleep in order to respond at any hour or consume endless energy drinks to work or play through the night. Studies of micro-workers around the world or “gold-farmers” in China have shown how these activities result in repetitive strain injuries or severe eye problems.
But unlike the industrial workplace, digital workers’ geographic isolation from one another, and from their employer, constricts opportunities for collective action to resist these conditions. Despite this, there have been some creative efforts to fight back against the precarity and exploitation that define microwork. These include the “Dynamo” email campaign — a quasi-union — and the FairCrowdwork monitoring system designed by IG Metall in 2015.
Global Gold Farmers
The way that digital platforms leverage gender, class, and race disparities to extract unpaid/underpaid digital labor hasn’t been addressed by any of the development establishment’s campaigns. These questions, which have been highlighted by theorists like Casilli as a “digital decolonial turn,” are crucial to understanding the expansion of this kind of work into the Arab world.
The “m2Work” project seeks to overcome “geographic obstacles” in order to offer employment opportunities to local youth and women. But this emphasis on “overcoming” geography obscures from how uneven geographies of power and wealth play out in the field of global digital labor. A study by the Oxford Internet Institute based on sixty thousand anonymized transactions completed on Upwork revealed a clear neocolonial pattern as “tasks” are sold from India and the Philippines to the United States, Australia, or the UK. This results in the same toxic race to the bottom that has defined industrial production for the past thirty years.
It is significant that the World Bank thinks that Gaza and the West Bank are “particularly relevant” destinations for microwork. In the global race to the bottom, the “bottom” is an open-air prison and an area under military occupation. According to their 2014 report, these sites pose “limited risks” to employers in terms of having to pay any full-time employment benefits or a minimum wage. Protections may be “risky” for employers. But making someone dependent for their survival on an algorithm with zero accountability is risky. Not paying them an adequate hourly rate to feed themselves or their family is risky. Irregular or excessive working hours resulting in repetitive strain, eye problems, or insomnia are risky.
One of the big selling points of digital workplaces is also that they are blind to race and gender. But the inherent inequalities of these systems don’t go away just because participants remain anonymous. After Chinese labor camps were exposed for forcing prisoners to play profitable online games, gaming chatrooms began to use “Chinese gold-farmer” as a pejorative for worker-players trying to sell virtual goods. Syrian players in Shatila who viewed their gaming activities as leisure rather than work complained that their broken English or Arab names meant they were viewed by other players as unwanted guest workers or “gold-farmers.” Microworkers have no protection from similar dynamics on digital labor platforms.
The World Bank’s m2work borrows heavily from the Samasource initiative, a non-profit business that seeks to alleviate poverty by outsourcing digital work from Walmart, eBay, and other companies to impoverished populations. But contrary to their marketing, these schemes don’t work in spite of predicaments like the occupation or the refugee crisis — they work because of them.
Digital labor plays perfectly into the logic of the refugee camp or the blockaded territory wherein subjects must be kept alive, but only just. Now, even stateless persons stripped of their political existence, of the rights and protections of the law, can be made productive. Through online platforms, people in places like Gaza, the “world’s largest open air prison,” can literally work within their figurative jail cells. Their integration into the global economy as cheap outsourcers makes users dependent not only on Israel or on a host state but on technology as well.
Like the prison industrial complex that profits off the unpaid or underpaid labor of inmates, these initiatives thrive off social and political upheaval in the Middle East. In fact, prison labor programmes in the United States now also include digital data entry, proofreading, and document preparation, just like the Chinese inmates forced into “gold-farming.”
These unconcealed efforts to overcome local regulation in the Global South to serve the interests of Western platforms are part of a broader pattern of what developmentalists call “tech for good.” But tech is not good in and of itself. Tawil-Souri urges us to recognize the underside of “progress” represented by the internet’s role in Middle Eastern societies, which crucially includes “the further consolidation of corporate capital and the imposition of its criteria and priorities on the nation state.” For all their inflated rhetoric of providing “access to the global knowledge economy” these World Bank–recommended schemes will do nothing to address the underlying structural problems that have caused high rates of unemployment in the first place. In Palestine, for example, the World Bank’s policies since the nineties have prioritized private-sector growth as the path to peace and state-building, ignoring reports showing that if the occupation was lifted, GDP would double. “Tech for good” is a natural extension of this faulty logic.
The Fate of the “Turk”
As Golombia observed, the pipe-smoking “Turk” from which Amazon’s microwork platform gets its name was part of a craze for automata designed to resemble the oriental “Other,” the docile and obedient Muslim of Christian theology. But the Muslim-as-machine takes on new meanings as workers in impoverished areas, from Syrian refugee camps to the Palestinian occupied territories, are forced to perform these repetitive, unskilled tasks, concealed behind a slick, anonymized interface. Machine-like, always-on, this “surplus population” can always be tapped into by companies to fuel the twenty-four-hour business cycle that drives Western progress.