Uber and CrowdFlower are two links in a growing supply chain of services that use APIs and human computation to put people to work. Uber uses CrowdFlower's API to pay someone to review the results of Ayesha's work, and, if it passes muster, it will process Uber's payment to her within minutes. If it doesn't meet the preprogrammed bar, Ayesha won't get paid for her efforts, nor will she have any meaningful opportunitiy to launch a complaint. the API isn't designed to listen to Ayesha.

[...] most automated jobs still require humans to work around the clock, often part-time or on a contract basis, fine-tuning and caring for automated processes when the machines get stuck or break down, as technical systems, like humans, are apt to do.

It's also true that the long march toward automation has historically created new needs and different types of human labor to fill those needs. In this respect, the new, software-managed work world shares features of the factory jobs that assembled cars by placing workers on a production line where and when they were needed most. It also resembles the so-called piecework that women and children did on farms in the 19th century, assembling matchstick boxes for pennies a pop. And it overlaps in obvious ways with the outsourcing of medical transcription and call center work to the Global South that boomed with the expansion of the internet in the late 1990s.

As AI advances, it creates temporary labor markets for unforeseen and unpredictable types of tasks. The great paradox of automation is that the desire to eliminate human labor always generates new tasks for humans. What we call "the last mile" is the gap between what a person can do and what a computer can do. Without a doubt, software developers willl use ghost work to perform the tasks at hand and push AI to its limits. [...] dependency on temporary human labor has always been a part of the history of technology's long march toward automation. [...]

Indian citizens are the only international workers who can also earn cash for their MTurk ghost work. There's no reason for this other than the fact that Amazon's multinational holdings allow it to operate and transfer money between its U.S. and India office locations. India's MTurk workers could opt to fill up an Amazon gift card, thoguh the company doesn't reliably deliver to many of India's sprawling, informal neighborhoods. If a worker in India wants to move their money to a personal bank account, they must first hand over their birth date and a scanned copy of their permanent account number (PAN) card, the equivalent of a U.S. social security number. [...] Amazon can, for an additional fee, cut paper checks or offer direct deposits to India's MTurk workers.

[...] In 2012, a CrowdFlower worker, Christopher Otey, filed suit against the company [...] 19,992 CrowdFlower workers had signed on to the lawsuit [...] CrowdFlower set expectations accordant with a full-time employee, but the pay and benefits were commensurate with those of an independent contractor. [...] "I didn't have control over the work I did. It was all done on their platform. I couldn't choose my own hours. I had to work when they provided the work. They pretty much controlled all the aspects of the work that was being offered." Given the degree to which CrowdFlower set the terms of employment, argued Otey, CrowdFlower owed him and other workers minimum wage, per the Fair Labor Standards Act. The company's legal team countered that, because CrowdFlower's workers were "free contractors," the FLSA didn't apply. Ultimately, in 2015, CrowdFlower paid $585,507 to settle the lawsuit, which left the question of the employment status of its workers unanswered.

[...] The new asssembly line systems [...] factory owners took control of both the raw materials and the labor "supply chain". Individual workers no longer coordinated production with each other. They were told where to stand, what to do, and how to hand off their contributions to be incorporated into the final manufactured product. In this way, the assembly line did not introduce a new division of labor between humans and machines so much as it wrested away control over the pace at which a person worked and their ability to coordinate with or delete to others.

[...] The largest, most progressive union, the United Garment Workers, organized in 1891, tried to root out what it saw as the "menace of the outworkers" and make them "a coherent part of its growth," but to no avail. Union organizers focused on getting young women to fill vacancies on the factory floor. But these approaches did not contend with or even recognize how often women worked on contract through piecework, because factory work was still considered morally suspect for a young, unmarried woman and, practically speaking, took her away from her other full-time job of cooking cleaning, and caring for children and elders at home.

[...] union strategies did not prioritize or recognize the specific burdens or costs women faced leaving contract labor or home-based piecework behind. Unions quickly abandoned their focus on recruiting young women to the factory floor. And none imagined that advocating for gender equality in the home to reduce women's household workoad might be a necessary strategy for unionizing women in the workforce. Instead, they shifted their attention to blocking the uptake of newer technologies that sped up the work pace. [...]

News magnate William Randolph Hearst lost a case in 1944 before the Supreme Court when he failed to convince the court that paperboys were the equivalent of their own bosses, operating as contract workers, exempt from employment labor protections under the Fair Labor Standards Act (FLSA). It's worth noting that paperboys didn't just deliver the news. They served as the primary sales force for papers, and their jobs often put them in physical danger as they made their way through the thick of bustling street traffic. The Supreme Court read the spirit of the FLSA as applying to any worker economically dependent on and producing for another entity. The Heart Corporation, bruised from the Supreme Court's verdict recognizing street corner paperboys as workers deserving of workplace protections under FLSA, lobbied to narrow the definition of "employee" under the Taft-Hartley amendment to WAgner's National Labor Relations Act. Hearst pushed to explicitly exempt independent contractors, working off-site or considered peripheral to business operations. Congress's revisions to the original Wagner Act would instead require courts to use strict tests to classify "common law" employment status rather than assume that businesses hiring people to carry out work counted as workers deserving of fair labor practices and employment benefits.

The erosion of labor protections for workers seen as "unskilled" but working outside of unionized manual jobs moved like water through cracks, exploiting and capitalizing on society's assumptions about whose work needed protection and who was worth protecting. Buried within the exemptions of the FLSA and the Taft-Hartley Act are clues about mid-20th-century assumptions about which workforces neede dprotection from unemployment and what - or who - seemed safe from automation [...] the FLSA excluded volunteering, what we might think of today as the ubiquituous unpaid college internship. Volunteering was considered apprenticing, core to building one's professional identity. Since medieval times, divinity, medicine, and law, teh so-called "learned professions," were seen as the exclusive paths of the educated classes. These skilled vocations needed no workplace protections. Their advanced degrees insulated them from economic insecurity. That is how a college education came to be seen as a gateway to the middle class for anyone who wanted to leave behind the mines or factories.

[...] The work conditions of the professions, at least in the early days of moder industrial capitalism, did not seem relevant to the debates over FLSA. And so, the Fair Labor Standards Act appied to "any individual employed by an employer" but not to independent contractors or volunteers training to enter the professional class. Both types of workers were considered liminal characters. The contractor was treated like a hammer or mechanical pencil, there only to do an immediate task. Interns, by contrast, were being groomed to step into the corner office someday.

[...] the 1980s, when temporary staffing agencies like Kelly Girl Service and Manpower contracted out more workers than most companies retained as full-time workers. Manpower's temp pool surpased GM's full-time employees in the 1990s. Sociologist Erin Hatton makes a compelling case that the wholesale swapping out of full-time employees for contingent staff was not just the fallout of technological change. She charts the staffing and temp agency sector's growth as its own service industry. What did it sell? The suggestion that young women and others, constrained by family care, limited educational opportunities, or geography, would reduce the companies' costs. Hatton argues that this "liability model," which treats workers as a drag on corporate profits, overtook the "asset model" of employment. [...]