Excerpted and modified from Uberland: How Algorithms Are Rewriting the Rules of Work by Alex Rosenblat. Out now from University of California Press.
Silicon Valley is famous for “disrupting” industries through innovative technology. But this language is also a coded way of saying that a new middleman is in town, with the implicit suggestion that the promising technology is inseparable from the business practices that become associated with it. This slippery rhetoric is part of how Silicon Valley companies create exceptions to the norms of their industry competitors and rewrite those industries’ rules.
Uber is one such new middleman. It has changed the rules of the ride-hailing industry by collecting and analyzing data from all of its app users and using this informational access to manipulate passengers and drivers alike through the platform. Uber has also taken advantage of its position as an intermediary to pocket differing commission than it agrees to collect from drivers on the fares passengers pay. And as Alison Griswold reported for Quartz, Uber’s “upfront” pricing helps the company overcharge passengers. But Uber may also be changing the definition of work by using the tools and rhetoric of technology to blur the line between entrepreneurship, employment, and consumption.
Uber brings the technology culture of Silicon Valley to the world of work. Facebook sparked a public outcry after it quietly experimented with the psychological states of select users by displaying happier or sadder posts to them in their news feed to study the effects of emotional contagion. People were outraged both because they didn’t want to be the unwitting subjects of mood experimentation, and also because the experiment contradicted the idea that a neutral, objective, and benevolent algorithm curates their news feed. Similarly, Uber experimented with driver pay by implementing upfront pricing without alerting drivers or adjusting their contracts, until months later, after drivers crowdsourced evidence of a new pay policy.
When Uber takes advantage of the unwitting users of its technology, it could be within its rights to do so, though its particular machinations actually contradict the company’s own description of its business model: In legal forums and in its contracts with drivers, the company says it provides a platform that connects all its users, implying that its technology is neutral, like a credit card processor. In one court hearing, Uber’s lawyers used rough metaphors to explain this logic in oral arguments, saying, “People demand ice cream. We have vendors, vendors who produce ice cream that are able, through our software, demanded—on demand to people that want ice cream. We facilitate that transaction. We’re not in the ice cream business, you know.”
But Uber is in the figurative ice cream business. Uber monitors drivers through the data they generate on the job and controls their workplace behavior through various methods, from in-app behavioral nudges that influence when and where drivers work to the threat of account deactivation if drivers don’t follow some of Uber’s behavioral “suggestions.” Yet Uber also explicitly adopts a model of customer service communications in managing its workers as if they were mere consumers. In fact, beyond intense supervision, Uber controls drivers by creating an appeals process that limits their ability to find resolutions to their concerns.
For years, drivers’ primary point of communication with Uber was by email. While Uber has added phone service and some in-person hubs in select locations, drivers still don’t have a dedicated human manager who responds to their inquiries. Instead, they have community support representatives at the email equivalent of a call center, often located abroad and managed by third-party companies. Effectively, Uber offshores and automates its main communications with drivers. Drivers receive automated replies to most of their inquiries, which often appear to be based on keywords in the text of their emails. In other words, Uber is managing drivers without a human that understands and is responsive to nuances. While automated responses might be practical for basic factual inquiries, they can prove woefully insufficient when a passenger overdoses in the backseat or harasses a driver. In my research, I have scrambled details of driver experiences to protect their anonymity, though I have made every effort to preserve the accuracy of their accounts. I have included the surnames of drivers who preferred to have their real names used. One driver, echoing a sentiment commonly posted in an online forum, remarked, “the support is nil at best. Most of the responses I get are less than adequate to the topic at hand. It seems that if it involves something other than a rider issue or a fare not starting on time, no one understands wtf you are talking about.”
Consumers are familiar with bad customer service, but this approach takes a greater toll in an employment context when drivers depend on community support representatives to resolve questions related to their livelihood. One driver I spoke with who drives for both Uber and Lyft in Atlanta told me a story about a passenger who accused him of drunk driving in the “passenger feedback” comments. He went to great lengths to explain to me that he works at night, when many passengers have been drinking, and that they can leave the smell of alcohol in the car. This driver, who is a diabetic, told me he would be passed out in the hospital if he had been drinking. When a passenger complained to Uber, he was instantly deactivated once the passenger submitted the comment, and he had to stop in the middle of working to write several emails to try to get reinstated. Another Uber driver, Jay Cradeur, authored an article on the Rideshare Guy detailing how he was deactivated for no reason, jeopardizing his livelihood, and then spent weeks trying to resolve the problem, without any success.
Even when Uber does offer a solution to an individual driver, it may not address the root cause or unfairness in the system, which creates a particularly challenging work environment. For example, one driver, Cecily McCall, from Pompano Beach, Florida, disputed a previously adjusted fare (where Uber claws back some of the driver’s earnings following a passenger complaint), writing in an email to a community support representative: “I had to end the trip early because the passenger got in the car, started to curse me out, [and] called me a dumb stupid nigga because I told him we was on an Uber pool. He said, ‘You dumb stupid niggas can’t ever get shit right.’ ” The community support representative replied with generic Uber policy details and robotic emotional drivel (“We’re sorry to hear about this. We appreciate you taking the time to contact us and share details.”) and emphasized, in bold print, that the passenger in question would not be matched with her as a driver again. Disgusted, the driver wrote back, “So that means the next person that picks him up … he will do the same, while the driver gets deactivated. Welcome to America.” Uber’s response also signals that McCall might be matched with other harassing passengers that drivers have reported, just not this particular one.
Because community support representatives are responsible for mediating disputes between Uber’s drivers and Uber’s passengers, low-quality responses multiply the unfairness that drivers experience at work, particularly because Uber’s technologically mediated employment is still plagued by classic workplace problems—like sexual harassment. For example, Leticia Alcala shared with me experiences she had of sexual harassment by passengers. She said she reports harassment to Uber, and the company promises not to match her with those passengers again. “They may not match you with them, but they’ll match other women with them,” she fumed. And some drivers fear recrimination if they detail an issue like harassment to community support representatives. Another driver who told me that he has been sexually harassed on the job multiple times fears that if he reports such an incident, the passenger may turn the accusation on him. He could lose access to the Uber platform and his livelihood. Drivers know that Uber controls their access to future rides, and therefore drivers may just fold sexual harassment and other complaints into the cost of doing business.
Community support representatives also serve as a poor replacement for driver management when it comes to the wage discrepancies drivers have experienced. In forums, many drivers comment on unpaid or missing wages that they tried, and failed, to collect, such as cancellation fees. Many drivers check their pay stubs weekly or monthly, while others don’t look at them carefully until tax season. Let’s say that a small percentage of drivers are actually going over their pay stubs each night and manually tracking their wait times. Only a percentage of those are going to write to Uber to complain. If a driver is willing to go back and forth with community support representatives, she might get her cancellation fee. Many drivers have demanded satisfaction from Uber and some showcased proof of their hard-won wages in forum postings. But given the time that the accounting and communications process would take, many drivers decide it’s not worth it for what could amount to a little less than $4. Drivers figure they’ll probably make more money by spending the same amount of time on doing an additional trip instead. Some say they give up after three to six emails with unfeeling community support representatives because it’s simply not worth the pocket change they might recover.
The customer-service-as-management process essentially provides disincentives to drivers to collect the wages they’re owed. This practice is analogous to cellphone companies cramming small, unauthorized fees into customer bills. Only a low percentage of customers actively track their bills, and only a low percentage of those are willing to spend an hour on the phone with a well-meaning but ineffective customer service agent to get a refund. Effectively, in app-mediated work it’s possible to withhold small amounts from wages (“fees” in Uber parlance) on a massive scale. Contesting this kind of wage omission doesn’t make sense on an individual level—it makes more sense to address it as a systemic issue on behalf of a large constituency of drivers. But the fact that drivers have to contest missing wages through community support representatives at all tells us how Uber speaks to the combination of consumer logic and employment logic of Uber’s model. The role of community support representatives substantiates the idea that Uber’s drivers are consumers of technology rather than workers.
The very vocabulary that Uber deploys to describe its drivers and its own practices reinforces this view of labor: It treats its workers as “end users” and “customers” of its software. The terms are used in Uber’s lawsuits, and a senior Uber employee casually referred to the company’s workforce as “end users” in conversation with me. The rhetorical impact of that language is clever. By fudging the terms of employment within its control, Uber provides us with a template for questioning what we know about employment relationships that can create legal distance between a worker and an employer. And it ushers in a new way of doing business all while the same old problems, like workplace harassment, persist under the veneer of technological neutrality.
Uber has tried to maneuver around legal walls using this thin argument of technological exceptionalism to protect its management practices. However, Judge Edward M. Chen seemed to find the company’s reasoning highly improbable in a class-action lawsuit brought by Uber drivers alleging that Uber violates labor law by misclassifying drivers as independent contractors, rather than as employees. When presented with the idea that drivers are customers of Uber’s technology, he said, “The fact that you screen drivers, select them, the fact that you, Uber, sets the fare, not the drivers, the fact that the company could not operate and exist as a company and make money without drivers, you think that does not establish, among other things, that these drivers serve Uber?” Uber’s shifts between the language of labor and the language of consumers evoke its earlier tactics of regulatory arbitrage. There’s no “sharing” in the sharing economy it has come to represent. In practice, drivers are hardly the entrepreneurs that Uber implies they are, or even true partners with Uber, even though the company calls them “Uber Driver-Partners”; drivers are not suspended or fired, they are “deactivated.” This conflation of workers with customers is clearly cause for disbelief. And yet, the miscategorization has deep roots within Uber’s claims about the employment relationship it has with its drivers. Regulators may support that blurring by using language consistent with Uber’s own: In 2016, the Federal Trade Commission brought legal action against Uber on the basis that it had misled drivers about their earnings, but the FTC also referred to Uber drivers as “entrepreneurial consumers.”
Uberland: How Algorithms Are Rewriting the Rules of Work
By Alex Rosenblat. University of California Press.
Slate has relationships with various online retailers. If you buy something through our links, Slate may earn an affiliate commission. We update links when possible, but note that deals can expire and all prices are subject to change. All prices were up to date at the time of publication.
The implications are stark: If the problems that Uber drivers experience at work can be reframed as customer satisfaction problems, these drivers lose access to remedies like employment law, which is available to workers in other businesses to redress any harms they suffer as part of their employment. And it isn’t just Uber using this language—it’s echoed by other companies, including Lyft and, across the ocean, a British food delivery service called Deliveroo. The kind of employment relationship that Uber has with its drivers is not unique. Rather, it signals a greater social force that is turning workers into customers through the power of technological tools and narrative. Deliveroo, for example, classifies its food-delivery people as independent contractors in a dozen or so of the countries where it operates, and classifies some, in other countries, as employees, such as in the United Arab Emirates. According to several Deliveroo employees I spoke with in 2017, this boils down largely to the laws that define employment relationships in those respective places. But the company contorts itself in order to avoid giving the impression that any of its workers are considered employees. As the Verge reported in April 2017 about a leaked Deliveroo document: “It says bicycle couriers who work for Deliveroo are never to be referred to as workers, employees, or staff, and that the Deliveroo jackets they have to wear on the job are not uniforms but ‘branded clothing.’ These workers don’t have ‘contracts,’ says the document, but ‘supplier agreements.’ They don’t ‘schedule shifts’, but ‘indicate their availability.’ ”
The central conflict of how to categorize a driver—and how to consider work in the sharing economy more broadly—animates the conflict between labor advocates and Uber. And Uber’s defense of their labor practices articulate dynamic changes in how employment and consumption are negotiated in digital spaces. The question in this new economy is whether algorithmic management really creates a qualitative distinction between work and consumption. Because by encouraging this distinction and describing its technology as a way to merely connect two groups of users, Uber can have its cake and eat it too, avoiding responsibility for prospective labor law violations while its ostensibly neutral algorithms give the company vast leverage over how drivers do their work.
Future Tense is a partnership of Slate, New America, and Arizona State University that examines emerging technologies, public policy, and society.