Advocates, academics, and policymakers alike have increasingly raised digital manipulation — the attempt to influence digital users’ behaviours and decision-making — as a cause for concern. This problem is primarily discussed in the context of consumers, such as with the Cambridge Analytica scandal, or with personalised advertising to teens on Instagram.¹ But this manipulation does not stop at the consumer, but also affects workers, whose employees increasingly subject them to an array of digital management techniques at work. A particularly salient example of this is Uber, who have — perhaps more than any other company — leveraged such digital manipulation as a labour management technique. Because Uber insists that its workers are not employees but independent contractors (a view reflected in labour regulation throughout the United States), it cannot directly control drivers’ schedules or routes. Instead, the company has devised a range of techniques, borrowing from behavioural sciences, to covertly manipulate drivers into working at certain times and in certain areas.² These include tactics such as sending drivers carefully crafted texts and pop-ups to keep them on the road; automatically queuing rides; or sending drivers push notifications that attempt to convince them to keep working whenever they try to log off.

Critiques of Uber’s labour practices are widespread. Often, scholars and advocates have grounded their arguments against these practices in appeals to drivers’ autonomy.³ While self-determination is always constrained under waged labour, app-based employment holds the promise of expanded freedom and choice, and many choose to drive for Uber because it lets them decide when and where they work. But Uber’s use of digital manipulation, or so the argument goes, diminishes individuals’ ability to freely make these decisions. However, using the framework of autonomy to understand how Uber uses digital manipulation limits how such techniques of manipulation can be understood — and challenged. Instead, Uber’s practices are better described through the theory of alienation.

In a recent, widely-cited legal paper, Daniel Susser, Beate Roessler, and Helen Nissenbaum argue that Uber’s digital management techniques are harmful because they degrade workers’ autonomy. The authors define autonomy as “the capacity to make one’s own choices, with respect to both existential and everyday decisions.”⁴ They posit that individuals can, for the most part, rationally deliberate and act according to the reasons they think are best, and that digital manipulation subverts this individual decision-making power, thus undermining autonomy. This degradation of autonomy is of grave concern to the authors. To them — due to the relationship between independent decision-making and democratic institutions — autonomy lies at the very normative core of liberal democracy. In their understanding, then, autonomy takes the role of a necessary background condition to well-functioning liberal society.

To understand this conceptualisation of autonomy and its relationship to digital manipulation, we must examine the intellectual tradition from which they draw. Indeed, the principle of autonomy is a cornerstone of traditional liberal thought. John Stuart Mill, one of the most influential liberal philosophers, describes autonomy as essential to human flourishing, a key element of personal well-being. In his so-called liberty principle, he argues that an individual’s “self-regarding thoughts and actions” ought to be protected from interference. Mill’s critique was aimed at state interference rather than at the private sector; but his liberty principle certainly echoes autonomy-based critiques of Uber’s digital manipulation, which argue that such practices interfere with individuals’ decision-making and therefore undermine self-determination and well-being.

The autonomy argument is convincing — in large part because it reflects the documented experiences of Uber drivers. Indeed, in legal scholar Veena Dubal’s ethnographic work with Uber and Lyft drivers, she finds that drivers frequently report a lack of control and autonomy.⁵ As one driver explains: “It really feels like you are being manipulated [by Uber]… it literally feels like you’re being punished by some unknown spiteful god.” This experience of manipulation is at odds with the promises of app-based work. The classification of drivers as independent contractors means, according to the U.S. Internal Revenue Service, that they “have the right to control or direct the result of their work”.⁶ In practice, however, they report a very different experience. As another explained to Dubal, “It’s like being gaslit every day, being told you are independent and being manipulated in all these different ways. Every single day, they are figuring out how to exploit you in different ways.”

---

Despite capturing some aspects of drivers’ affective experiences, the autonomy-based critique has severe limitations that become clear when Uber’s business model is examined more closely.

By identifying unfreedom as the primary harm caused by digital manipulation, Susser et al. presuppose that Uber drivers would enjoy freedom if they were not subject to these practices. Without Uber’s use of digital manipulation, it is implied, drivers could act as entrepreneurs with the capacity to make their own choices. This assumption is, of course, starkly at odds with the experience of waged labour under capitalism, in which a worker — having nothing to sell but their labour-power — is entirely dependent upon doing so to survive. The focus on unfreedom is also entirely at odds with how Uber’s business functions. Digital manipulation is not simply a technique that Uber occasionally uses to adjust drivers’ behaviour; it is, rather, central to how the company secures profits. Uber’s very profit-model, that is to say, relies upon the ability to render drivers de facto employees, by exerting control over where, when, and for how much they work — all the while evading the financial and legal responsibilities of direct employment.

In the U.S., rideshare companies have worked tirelessly to ensure that their drivers are classified either as independent contractors or as “third-category” workers, an employment status that falls between employee and independent contractor. This allows Uber and other companies to exert power similar to that of an employer (such as by slowing down rides offered or locking drivers out of the app to effectively control their work time), while shifting employment costs and responsibilities (like minimum wages, paid leave, and benefits) onto workers.

U.S. employers have long used race to justify differential worker rights. For example, the Fair Labor Standards Act initially excluded domestic and farm workers — professions dominated by Black and immigrant workers — from minimum wage protections. This carve-out, which lasted for decades, essentially legalised lower pay for racialised sectors of the economy. It was only following persistent pressure from social and labour movements that Congress amended the FLSA, in the late 1960s. Uber continues this legacy of treating racialised workforces as second-class employees, who are controlled like actual employees but receive none of the protections.⁷ Subal notes that the rideshare workforce is made up primarily of immigrants and people of colour. “But rather than addressing racial inequalities by improving the precarious working conditions of their primarily people-of-colour workforce”, she argues, “the rideshare companies Uber and Lyft have used the existence of these inequalities as a resource to justify and legalise their business model” — via independent worker classification, that is to say. As was the case with agricultural and domestic workers, the racial makeup of the rideshare workforce has everything to do with their (mis)classification as independent contractors — a classification that constitutes yet another chapter in U.S. labour law’s history of racial exclusion.

---

Rather than autonomy, alienation offers a more accurate and politically potent framework for understanding the digital manipulation of labour, helping to reveal how this practice threatens not just drivers’ autonomy, but also their economic security, and their capacity to see themselves as part of a collective. In Marxist theory, alienation describes a structural, objective condition under a particular political-economic system — capitalism, that is — rather than simply being a subjective experience. In the Grundrisse, Marx describes how, under capitalist production, working people are separated from both the process and product of their labour. This separation has profound consequences. As workers, we must sell our labour-power to access the material basis necessary for survival. We are “doubly free”: free to sell our labour-power, and free to otherwise starve. Those who own the means of production dispossess us of the products of our own labour, transforming such products into commodities, the sale of which yields more capital. Our labour, therefore, results in an ever-increasing productive power for capitalists. Meanwhile, our daily life as workers holds no relation to our desires, no relation to our self-expression, and no relation to who we are or might try to become — it is alien work. And with such work as the central organising principle of capitalist society, we become alienated from ourselves and from others.

Unlike autonomy, alienation foregrounds the material relations between the worker and the owner of the means of production. Whereas autonomy focuses solely on an individual worker’s affective experience, alienation links it to workers’ collective exploitation, and to capitalists’ accumulation of wealth. Consider, for instance, Uber’s use of personalised wages — one of its most powerful digital manipulation techniques. Over the past five years, Uber and other rideshare companies have begun to use driver, consumer, and other contextual data to generate targeted payment offers calculated by means of a highly opaque algorithm. An experiment by media outlet More Perfect Union showed how Uber offered drivers in the exact same location different rates for the same rides — proving that rates are, at least in part, calculated using individual drivers’ behavioural data.⁸ This practice is enabled by the surveillance and legal infrastructure that surrounds gig work. Since rideshare drivers are not considered employees, Uber does not need to comply with minimum wage laws. Further, both drivers and passengers are subject to the tracking of their location, their transactions, and their behavioural patterns. This information allows Uber to calculate a personalised wage, which is essentially the lowest possible payment that they can get a particular driver at a particular moment in time to accept. As Dubal explains: “individual workers are paid different hourly wages — calculated with ever-changing formulas using granular data […] — for broadly similar work.”⁹

Like Uber’s other digital manipulation techniques, personalised wages are hidden from drivers, an opacity that is arguably intentional. Black-box pay algorithms make it more difficult for drivers and regulators to hold companies accountable to fair and transparent pay standards, while also enabling Uber to adaptively manipulate drivers’ behaviour, by identifying the lowest rate at which a driver will still accept a ride.¹⁰ While the precise formula remains hidden, the purpose is clear: to exploit a driver’s vulnerabilities and incentivise behaviour that benefits the company.

Seen through the lens of autonomy, Uber’s use of personalised wages to influence driver behaviour is problematic primarily because it threatens a driver’s decision-making power. The political economy of this practice, however, is beyond the scope of critique. But clearly, personalised wages are first and foremost a material practice that results in lower wages for the collective body of drivers, and higher profits for Uber. As scholar Zephyr Teachout has argued, these personalised wages function as a tool for wealth transfer.¹¹ The resulting wages for drivers are abysmal. One study found that after expenses, drivers take home an average of $6.20 per hour.¹² Uber’s CEO Dara Khosrowshahi, meanwhile, earns nearly $40 million every year.¹³

A framework of analysis based on alienation helps to explain how personalised wages continuously reinforce Uber’s ability to extract wealth from drivers. The separation of workers from the product and process of their labour means that the more they work, the more surplus-value they generate for capitalists. The more a driver works, then, accepting personalised payment rates for rides, the more data Uber can collect, and the more it can finetune its wage-targeting systems. This data includes not just ride transactions and ratings, but also everything from how quickly a driver brakes to how frequently they stop, and for what and where. This information populates driver profiles, which the company can use to target wages to match perceived driver incentives.¹⁴ Thus, the harder drivers work, the more personalised their wages become; and the more personalised their wages become, the harder they must work. As Dubal found in her ethnographic work, the longer drivers worked, the lower their hourly wages would fall. Personalised wages, Uber’s profit, and the impoverishment of the workers are thus recursively linked.

Within a framework that only recognises the individual’s loss of power, we are also unable to see how digital manipulation functions as a tool of de-collectivisation — constraining the ability of workers to see themselves as part of something larger, and to construct a collective form of autonomy. Dubal describes how Uber drivers often notice that they earn different amounts than their peers, even when they drive roughly the same routes and hours. These differences in pay generate feelings of individual failure and shame. But they also work to atomise workers, corroding the social ties on which collectivisation and organised resistance depend. As one driver told Dubal: “Any time there’s some big shot getting high payouts, they always shame everyone else and say you don’t know how to use the app.”

Personalised pay pits driver against driver — the war of all against all. It thwarts efforts to build solidarity and organise. When workers are separated from the process of labour and thrust into algorithmic silos, they can only relate to other workers as adversaries, or tools in their pursuit of wages. Their capacity to relate to one another as potential collaborators is inhibited, and their efforts to organise are thus stymied. This group-level effect is not simply a composite of individual harms, but a result of something inherently relational.

Ultimately, drivers continue to agitate and organise — as have the generations of workers that have come before them, confronted with the technological impositions of their own era.¹⁵ As Sergio Bologna once stated: “Every attempt to dissolve class identity through fragmentation, individualisation, and dispersion ends by producing new subterranean forms of collective behavior, invisible until they suddenly erupt.” What form this resistance will take might be difficult to anticipate in our present moment — but if the history of class struggle tells us anything, it is that it is inevitable.

Notes

1. Nicholas Confessore, “Cambridge Analytica and Facebook: The Scandal and the Fallout So Far”, The New York Times, 4 April 2018. [^]

2. Noam Scheiber, “How Uber Uses Psychological Tricks to Push Its Drivers’ Buttons”, The New York Times, 2 April 2017. [^]

3. Isaac Chotiner, “When Your Boss Is an Algorithm”, Slate, 26 October 2018; Sarah Kessler, “How Uber Manages Drivers Without Technically Managing Drivers”, Fast Company, 9 August 2016. [^]

4. Daniel Susser, Beate Roessler and Helen Nissenbaum, “Online Manipulation: Hidden Influences in a Digital World”, Georgetown Law Technology Review, 2019. [^]

5. Veena Dubal, “On Algorithmic Wage Discrimination”, Columbia Law Review, 2023. [^]

6. See: https://www.irs.gov/businesses/small-businesses-self-employed/independent-contractor-defined. [^]

7. Veena Dubal, “The New Racial Wage Code”, Harvard Law & Policy Review, 2022. [^]

8. Eric Gardner, “Here’s What Happened When We Put 7 Uber Drivers In the Same Room”, More Perfect Union, 3 April 2024. [^]

9. Dubal, “On Algorithmic Wage Discrimination”. [^]

10. Dara Kerr, “Secretive Algorithm Will Now Determine Uber Driver Pay in Many Cities”, The Markup, 1 March 2022. [^]

11. Zephyr Teachout, “Algorithmic Personalized Wages”, Politics & Society, 2023. [^]

12. Eliza McCullough, “Prop 22 Provides Drivers with Inferior Benefits to Those Guaranteed to Employees”, 2022. [^]

13. See: https://aflcio.org/paywatch/UBER [^]

14. Teachout, “Algorithmic Personalized Wages”. [^]

15. For example, see this recent strike led by app-based drivers in the UK: Schannell Kanyora, “Inside private hire drivers’ strike: 18 hour shifts, passenger violence and unfair pay”, The Mirror, 14 Feb 2026. [^]