Resisting The Algo-Boss
The Lies That the Apostles of Neoliberal Capitalism Tell Us
Despite being told by the apostles of neoliberal capitalism that technology is neutral, natural, and even non-political, emancipatory politics must always destroy the appearance of the so-called “natural order.” This also applies to the rise of algorithmic management.
Neither traditional management nor algorithmic management are parts of some sort of natural order. Instead, they are a part of a capitalist order.
Particularly, the advent of algorithmic management and your all-new algo-boss might indicate that the future of work may not unfold in our favour. This is—at least in parts—because of the fact that algorithmic management and the algo-boss are not much more than a new way of organising work.
With the algo-boss on the horizon, workers are increasingly directed, watched, monitored, controlled, tracked, measured, and meticulously assessed by a computer system that is based on an algorithm.
While the use of tools and machinery in companies is not new, the introduction of new machines and new work systems have not always replaced workers.
However, it is equally true that throughout the history of capitalism, human workers often remained cheaper, more flexible, more reliable and easier to replace with other workers than introducing expensive, high-maintenance, frequently malfunctioning, and rather inflexible machinery and robots. Meanwhile, it is also the case that algorithm-based robots can do things that people simply cannot.
For example, a Rubik’s Cube has 43 trillion possible options. The fastest person needs 2.68 seconds to complete the cube. A robot does it in 0.38 seconds. In other words, in so-called closed systems with confined parameters, artificial intelligence and robots can beat every human—every time.
But it is not just speed that attracts management to algorithms. Algorithms become a handy ideological tool when the algo-boss needs to legitimise itself, their company and—ultimately—also capitalism. This is because algorithmic system adds an air of infallibility and objectivity to management.
Both are something management never had, does not have and is unlikely to ever have. Hence, the use of algorithms can give management and the new algo-boss a new façade. It is this additional ideological support that renders algorithmic systems so important to management.
On the upswing, algorithmic management also means that—in most cases—workers are being managed by computers rather than replaced by them. In other words, it is middle-management that is more likely to be replaced by algorithmic management rather than workers (down) or top-management and CEOs (up).
This too, is made to appear as if all this has nothing to do with capitalism, companies, or your all- new algo-boss who, in fact, are doing this.
While mathematical algorithms are made to appear rather technical—just as machine learning, programming languages like Python and Java, etc.—the introduction of algorithmic systems into companies is done by real people (read: management).
The introduction of algorithms does not happen in a vacuum. Instead, algorithmic systems are inextricably linked to both algorithmic power and corporate power. Beyond all this still lurks the power asymmetry between workers and management.
Consequently, algorithmic management is a mode of organizational politics and that, in turn, is linked to the frontier of control that constantly moves between managers and workers.
Perhaps the most important element behind the algo-boss is not the much celebrated surveillance of workers. Instead, the principal dogma governing corporate management is to minimise waste, to maximise efficiency and profits. This is done for the sole purpose of cranking up profits.
This is the raison d’etre (the reason why they exist) that companies and management are fond of algorithmic management. It promises efficiency, the reduction of waste, and increased profits.
It might come as a shock to devotees of surveillance and Foucault’s panopticon, but algorithmic management is not about management becoming Big Brother to spy on workers. Instead, it is all about the profits which drives the algo-boss.
In the all-embracing code of algorithmic management, the entire labour force of a company is regarded as a form of “waste.” And—so the mythology of management goes—this waste needs to be eliminated. In other words, beneath the ideological façade of flexibility, personal autonomy, and “self-employment lurks managerial despotism.
One signifier of all this has been the fact that workers in algorithmic-driven warehouses, for example, have been forced to urinate in bottles.
This came as a no surprise since algorithms are about the allocation of work tasks under the dogma “no waste—no time.” The corporate fallacy of “time is money” becomes guided by an algorithm. At the surface, it appears as if an algorithmic system forces workers to pee in a Coke bottle.
Yet, it is not the algorithm that forces workers to pee in bottles, but it is “your all-new algo-boss” who is now armed with algorithms that does it.
With the Age of Machinery rapidly becoming the Age of Algorithms, algorithmic management transformed workers even more efficiently into mere tools that are at the disposal of “their” computational system. With that, a rather unassuming technology like artificial intelligencehas enormous so-called “practical” effects: efficiency for management and “bottle-peeing-style” inhumanities for workers.
The truth behind all of this is that the entire system is practically a science of making money. And for that, and under algorithmic management, workers have been reduced to information-generating appendices to algorithmic systems.
To semi-regulate work under algorithms, a British trade union called IWGB has signed a so-called “partnership deal” with an online platform. It was the first union agreement that Uberhad entered into. On the downside, it was one of the largest union-platform deals signed by a union “without” a mechanism for collective bargaining on pay, let alone Uber’s algorithm regime.
Yet, it can be argued that such a recognition deal can be used as a way for algo-bosses to ward off more successful or more forceful trade unions.
Meanwhile at a British supermarket called Asda, its distribution arm has a rather remarkable unionisation density. This has been achieved by a British trade union called GMB which covers around 83% of the workforce.
However, there has also been a shift in trade unions’ strategy. Rather than focusing on workers with plans for direct industrial actions and strikes, British unions have increasingly focused their resources on initiating and engaging in political lobbying and judicial court proceedings. On the whole, these kinds of “solutions” do not involve workers directly.
At Amazon’s warehouses, for example, work is not really organised and run by managers as such. What makes modern logistics so effective is that—at crucial points of the operations—the system can organise and run itself. In other words, there is next to no need for middle-management, or for line supervisors and for the old-style shop floor overseer.
Instead of a direct supervisor, now it is the feedback loop that exerts control within a system run by algorithmic management. This has moved management from a human manager towards cybernetics.
At this point, the remaining managers are reduced to a servo-mechanical role. Put simply, some managers will remain because they are there to serve the machine—an algorithm.
In other words, algorithmic management demands relatively small contributions from a human manager. For workers and less so for managers, this comes with the rise of the algorithmic panopticon.
Behind all this are the words of Karl Marx who once noted that many of the technical inventions that have been made were done for the exclusive purpose of providing capital with weapons against a working-class revolt. As such, many algorithmic systems are invented to suit capitalist production.
Such inventions further bolster the domination of machines, robots, algorithms, and ultimately also the power of the algo-boss over workers.
Much of this is cloaked by management, business schools, corporate think tanks, and the adjacent business press through a targeted strategy of depoliticising technology.
Through depoliticising algorithms, the apostles of neo-liberalism can give technologies such as algorithmic systems the appearance as if they had nothing to do with relations of power.
This is done because the corporate press, think tanks, neoliberal economists, managers, CEOs, business school professors, and the adjacent neoliberal entourage know very well that much of this is about power rather than rationality.
To hide profits and capitalism. it is their job to make algorithmic systems appear rational and neutral. This is also done to prevent heretic thoughts that algorithmic systems might be no more than a self-activating monster that turbo-charges the domination of capital over workers.
On the upswing, a machinery like algorithms can also be a weapon. It can be stolen, captured, and sabotaged. And it can be used against workers or “by” workers.
Looking at any Amazon warehouse, for example, the totalitarian tendencies of Taylorismbecome visible. Spiced up with algorithms, the algo-boss can use algorithms against workers.
Yet, the corporate management, business school professors, and the pro-business press have an exceptional capacity to obfuscate power relationships. To engineer such system-stabilising ideologies, algorithmic management—whether despotic or not—is marketed as purely “data-driven.”
The idea is to make algorithms appear as it they have nothing to do with capitalism and profits. At the same time, algorithms also function as a kind of “social engineering software” setting in motion a computerised version of Blake’s Dark Satanic Mills.
In Amazon’s Satanic Mills—just as under algorithmic management—there is also the profit logic that is cranked up by establishing algorithmic discipline. In algorithmic disciplining, the overseer’s whip (18th century and 19th century) and the infamous three-strike-rule (20thcentury) has now moved inside an algorithm. It becomes less visible.
Under algorithmic management, those who fail to perform—as defined unilaterally by algorithmic management—will be punished by an invisible algorithmic system. Worse, performance goals are deliberately set on a higher and higher scale by the algo-boss. The nasty intention behind this is that workers will be forced to strive towards achieving those—mostly unachievable—goals.
In that way, so management believes, workers will be motivated to get more productive as performance targets move higher and higher. In other words, performance driven by algorithmic management means that workers are put in a perpetual state of underachievement. What awaits workers is an eternal rat race towards unachievable performance goals.
The entire setup is supported by an excessive belief of the algo-boss in the almighty power of algorithms. At this point, algorithmic management might even start to believe its own ideology.
While sticking to management’s time-honoured “command and control” formula, the belief in algorithms is communicated downward. Managers call this “cascading down.” A General Motor’s worker once said, this means, “they piss on us.”
Algorithmic management does not change the fundamentals of the information and communication asymmetry that defines virtually all managerial regimes.
While communication between workers occurs horizontally—the same level—this is now mediated through data-driven communication that is increasingly conveyed vertically through the algo-boss. In short, algorithmic systems strengthens the vertical—top-down—communication of the algo-boss.
Part of this is the AI-driven machine communication that shapes workers and management while reducing workers to being transformed into tracked and “censored” entities—for management. Information and data are passed into a system that is largely out of reach of a worker’s control.
Overall, workers can do little to intervene directly in its use by the algo-boss. It fortifies not just algorithmic authority, but also algorithmic authoritarianism.
Worse, the all new algo-bosses are able to distance themselves from shop floor workers. The algo-boss becomes a “black-box.” Algo-bosses and algorithmic management are made unaccountable. By creating distance, managerial despotism might get even worse.
Ever since Milgram’s experiments on obedience, we know that social or corporate distancemakes it easier to torment victims.
In other words, the algo-boss might be an even harsher form of management compared to traditional management as the distance between the algo-boss and workers becomes more distant.
Given what we know from Milgram’s experiments, there is the potential for an even crueller form of discipline by the algo-boss. In other words, algo-bosses can become more punitive given their ever-greater distance to workers as algorithms move between the traditional worker-manager relationship. All this comes, of course, with algorithmic management’s synthesis of panopticism and Taylorism.
Set against all this is the resistance through subversion. One might also argue that the simple act of “refusal” can mark the start of liberatory politics. Perhaps even more important is that “refusal” also symbolises a point of departure from the capitalist logic of profit making.
This is something that management and even more so, the apostles of the more ideological Managerialism seek to prevent—most desperately.
Beyond “refusal” are wildcat strikes. In the case of online platforms, these are often organised through WhatsApp groups. On the downside, such union-based WhatsApp groups can and have in fact been infiltrated by managers and adjacent corporate stooges.
To put it differently, Orwell’s Big Brother not only lurks inside warehouses but also inside your very own trade unions’ WhatsApp chat group.
That too, is not new as Henry Ford already employed company spies against the trade union. Notwithstanding, workers’ acts of refusal and resistance against algorithmic management often occur outside of union-sanctioned actions.
Yet, these acts of refusal are still “individual” ways of dealing with the corporate system of capitalism. However, this form of protest nevertheless lacks the experience of generating collective feelings among workers. Only collective action and strikes can generate a collective identity that can match corporate power.
In contrast, individualistic acts of resistance can start by simply doing what you’re not supposed to do. As one worker said, you take advantage of the lack of managerial manpower on the shop floor.
In any case, and under the conditions of the algo-boss, refusal and resistance has to operate under the algorithmic radar. In other words, algorithmic management and the algo-boss are something that workers can defy.