In the previous post, I considered the depersonalization of authority, and the role of judgment and deliberation, in algorithmic workplaces. To end this series, I want to turn to the place of resistance to algorithmic management.
Any account of the capacities of democratic self-direction will include the skills to assert your interests and the interests of others, to scrutinize and challenge decisions you think are unfair, and to resist attempts to be disempowered. These are skills that can be exercised individually or collectively, and for all the worries about the socializing effect of traditional wage labor, the workplace has historically been one of the more important training grounds for these capacities since before the rise of mass electoral politics. However little employers like it, workers will try to resist their bosses when they can, contesting either bad pay or poor working conditions, the more basic power structures that produce those outcomes, or both. And because workers are individually weak compared to their employers, there have always been great incentives for resistance at work to be collective.
The organized labor movement emerged from these workplace struggles, and in various forms that movement produced enormous political change, electoral and otherwise. Participation in this kind of collective resistance functioned as a ‘seedbed of the civic virtues,’ in the typically republican language of the US labor movement. Even for those socialists and communists who worried that this trade union activity would degenerate into mere ‘trade union consciousness,’ fighting only for a better place within capitalism rather than for deeper political transformation, the collective forms of resistance incentivized by the traditional workplace were at least a necessary first step in producing political subjects capable of organizing together to press political demands.
The decline of unions can’t be blamed on algorithmic management—it long predates it—and for a variety of reasons, including suffocating anti-union legislation in many jurisdictions, the relationship between the rank-and-file union member and the decisions of union leadership has been growing more distant for quite some time. But different kinds of workplaces make some kinds of resistance easier and some kinds harder. One of the things algorithmic workplaces make much harder is individual contestation within the firm, for the reasons I’ve already discussed: these algorithms are opaque, and there’s no obvious locus of decision-making to challenge, nobody at whom an individual worker can point the finger or from whom they can demand an explanation.
But another thing they seem to make harder is collective resistance itself. Workers tend to be isolated, and face huge epistemic barriers to learning enough about the algorithms that run their day to even frame their demands. The individualized monitoring that algorithmic tools allow means that workers are in competition with each other more, and more obviously, than ever before. And everywhere you look, you will find someone in a position of authority throwing their hands up and blaming ‘the system,’ always just above their pay grade, and always just beyond their understanding.
Although not strictly a consequence of algorithmic management, corporations like Amazon have also been credibly accused of using their repertoire of surveillance tools to monitor and crack down on union organizing. However true this may be, the opaque and pervasive digital surveillance of an algorithmic workplace certainly seems to make workers worry—reasonably—that their political activity is being monitored as well: a worker involved in the failed unionization campaign of a warehouse in Bessemer, Alabama, told a researcher about a ‘perception’ among employees that “everything is monitored and Amazon kind of knows everything, hears everything, sees everything.” This kind of paranoia (or, as the case may be, this kind of extensive surveillance) produces an extremely hostile environment for collective resistance.
This isn’t an all-or-nothing affair, and there are incredible examples of platform workers surmounting these odds and organizing (for example) wildcat food delivery strikes. And indeed, Amazon warehouses faced a wave of union organizing against terrible conditions over the last few years, with some important successes alongside the failures (some of these failures, as noted above, have been attributed to novel ways that algorithmic management tools were wielded as a union-busting measure by the corporation). But by far the most common form of resistance to algorithmic management is individual.
In some important senses, individual resistance is easier within a highly algorithmic workplace. Algorithmic monitoring has blind spots and glitches and, given the relative lack of human supervision, workers who find those blind spots can exploit them to push back on the demands of management. For instance, Todd, the supermarket picker, tells an interviewer that he and his colleagues discovered through trial and error that they could press the button marked ‘item too big’ on their scanning guns to pause the clock that calculated their work rate. They managed to claw back some breaks until the system was updated.
These tricks are often ingenious, a kind of asymmetric campaign of slipping through loopholes in managerial control. But the loopholes get closed, and they get closed more quickly the more people know about them. Todd mentions another trick, a code that only supervisors are supposed to know, that allows them to pause a worker’s clock. But managers leave that code lying around sometimes. He says that “[s]ome of the workers are private with that, because they know if only a few workers have that code it won’t [raise suspicions].” This is resistance, to be sure, but as an example of the kind of resistance to work incentivized by these management tools, it leans towards the private, piecemeal, the anti-associational. It doesn’t provide an opportunity for the kinds of public contestation, relationships of solidarity, and collective decision-making of traditional labor struggles.
All of this is exploratory. In some sense, only empirical work can really answer some of the questions I’ve been asking, and the deployment of these management tools will look different in different sectors. There might even be upsides, if we look carefully enough. But the rise of algorithmic management promises a significant change in most people’s experience of work, handing new and unfamiliar kinds of power to employers. If we’re to take seriously the relationship between what we do and who we are, then the coming transformation of the workplace should make us wonder how we, too, are being transformed in the process, in ways that we have reason to resist.
Editor’s note: This is an extra edition of the Perspectives on Democracy series for October 2026.
Orlando Lazar
Orlando Lazar is Lecturer in Philosophy at King's College London. His research focuses on work and the workplace, power and domination, and economic democracy. Recent work includes “Taking it Home With You: Work, Free Time, and Domination” (CRISPP, 2025) and “Micro-domination” (European Journal of Political Theory, 2023).
