- Research Programs
- Regions & Topics
- All Publications
A nation must think before it acts.
Mary L. Gray and Siddharth Suri, Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass, Houghton Mifflin Harcourt, 2019
Many of the apps we use—as smart and as fast as they appear—cannot function without people working behind the scenes to fill in the gaps. Machines are not as good as humans (yet) at figuring out whether a tweet is offensive or deciding whether an Uber driver’s authentication selfies match from one day to the next. But who are these hidden workers? Where do they live? New research from anthropologist Mary L. Gray and computer scientist Siddharth Suri unearths a key source of so-called “ghost work” outside the United States: India.
Is this form of labor a boon or a curse for the world’s largest democracy? For context, newly released data show that unemployment in India peaked at 6.1 percent in 2017-2018—a 45-year high. The unemployment rate was even higher for urban Indians (7.8 percent) and college graduates (13.2 percent)—and strikingly high for women with a bachelor’s degree or above (35 percent). Prime Minister Narendra Modi, recently reelected to a second term, will have another chance to make much-needed labor market reforms that prepare workers for the artificial intelligence era. In the meantime, higher-skilled workers, including many of the 600 million Indians under the age of 25, see few immediate opportunities.
Gray and Suri’s findings bring to life some of the peculiarities behind India’s unemployment statistics. To reach their conclusions, they first had to bypass application programming interfaces, or APIs, that render ghost workers invisible—often distinguishable only by the unique string of characters generated when they establish an account. The authors posted paid tasks on platforms like Amazon’s Mechanical Turk, or MTurk, where workers log on to complete on-demand work that businesses or any registered requester may post. They then collected data from workers who accepted the tasks through both surveys and ethnographic fieldwork.
They found that ghost work has been one avenue for those facing rising religious discrimination—which has surged in an era of Hindu nationalist politics under Modi—to find meaningful employment. Fareed, a devout Muslim living in Hyderabad, relies on paid work from Amazon’s MTurk platform to support his family. With few local employment options available, and uninterested in working as a hired driver in the United Arab Emirates, as many of his Muslim peers do, he also seeks to improve his English and job prospects.
Indian women who face scrutiny for working outside the home have also been drawn to non-discriminatory work they can do from behind a screen in the living room. “They could take on earning an income without throwing themselves into the tense national debates surrounding the impropriety of ‘call center girls’—women who work swing and night shifts, alongside men of all castes and religions, and who are often accused of prioritizing making money over propriety and piety,” write Gray and Suri.
Most of the non-U.S. workers the authors interviewed claim earning money is less important to them than gaining transferable skills. The authors met Anand, a 24 year-old student living in Chennai, who keeps a cheat sheet on the wall near his desk that enables him to quickly answer basic questions about the U.S. Anand’s MTurk income goes toward personal expenses, but he hopes it will one day lead to a position with Amazon.
Despite opening doors to workers with few other opportunities, on-demand labor leaves much to be desired. Ghost workers’ employment status is vague in both the U.S. and India, and they lack access to benefits and legal protections. Internet connectivity can be problematic in India, particularly during monsoon season, and is especially limiting in other parts of the developing world. Isolating, repetitive tasks without clear prospects for advancement take a psychological toll. The authors did find ample evidence of workers creating online communities where no infrastructure exists through Facebook groups that function like digital coworking spaces. But such coordination is ad hoc, and most ghost work platforms are not designed to accommodate deeper collaboration, let alone organizing.
The authors predict that these problems could have society-wide ramifications if left unaddressed. Gray and Suri cite a World Bank estimate that the professional on-demand labor market will grow to $25 billion by next year. By 2055, the authors estimate that 60 percent of today’s global employment will likely be converted into some form of ghost work. The “paradox of automation’s last mile” means there will always be a new frontier susceptible to automation, and people, unique in their creative capacities, will be needed to help reach it.
Workers, businesses, and governments—in India, the U.S., and elsewhere—must decide what they want ghost work to look like, argue Gray and Suri. They conclude the book with proposed “technical fixes for social change” that would make ghost work a more viable and humane form of long-term employment. If these fixes gain traction, perhaps on-demand work will attract more of India’s skilled jobseekers, while remaining a refuge for minorities who face discrimination in the traditional market.