The Everyday Resistance of Slacking Off at Work
Craig Gent examines resistance at work in the face of algorithmic management and digital control.
Strikes may be the primary tool in the arsenal of workplace resistance, but in algorithmic workplaces they are liable to be undermined by the very systems that govern the work. In response, just as the introduction of machines precipitated a need to reorientate working-class organisation from a mode based on workers’ technical knowledge and skills towards mass-membership trade unionism, under algorithmic management we can no longer rely on the forms of political mobilisation that have become our common sense.
To discover the political mode that is best suited to any organisation of work, we must understand work as a system, as an ecology, and determine where we can leverage power and control within it. Mass strikes are disruptive because they entail a large withdrawal of labour, but in the process there is also a withdrawal of workers’ contact with the productive process. But what if a work process was robust or smart enough to fill gaps in labour, or redirect work processes to other locations in real time, even if the barriers to organising strike action were lower than they are? Would we still view strikes as the culmination of workers’ power and agency?
If we want to win against the cyberbosses, we will need to think more broadly and creatively than today’s trade unions have done. If we can anticipate new directions in algorithmic management, then we also ought to precipitate new directions in workers’ political organisation. The natural starting point for this is to consider what actions workers are currently taking against the managerial forms that govern them at work, and to question how these actions can contribute to our understanding of contemporary class struggle in algorithmically organised workplaces.
For some people, any talk of subversion and subterfuge at work is fanciful. It is hard enough to recruit workers to a trade union, or to persuade them to vote for a legal strike. Yet it is no less fanciful than expecting mainstream union negotiators to understand the politics of algorithmic systems better than the bosses who patented them, especially when their own demands are stuck on the thorns of data transparency and AI ‘explainability’.
But why not just back unions, or seek to bring about change from inside them? Because the working class itself is more important than its unions as organisations. Trade union membership is a thin metric for understanding the extent of workplace resistance. It’s not that unions are unimportant; it’s that we should be unsentimental about them. Some unions sanction wildcat strikes or boisterous protests, others clamp down on them.
GMB was squeezed out of its efforts at Asos by Community after years of resources spent trying to organise workers there, only to do the same thing to IWGB at Deliveroo. Usdaw, a union specialising in representing retail workers, failed to leverage any power against supermarket employers at the time when its workers enjoyed the highest public support, while being subject to the deadliest public health crisis in memory. Large unions like GMB can take the position of being an auxiliary force assisting workers at Amazon while leading a superficial ‘partnership’ with Uber that effectively denies members a pay-bargaining mechanism. Sure, we can agree that unions are basically good, but there’s no political utility in getting misty-eyed about them. What I am interested in is the ability of the working class to struggle; I would like unions to help with that, but history tells us to be prepared for all eventualities.
During the Covid-19 pandemic many people, taking advantage of the technical reorganisation of life and work, found novel ways to skive or slack off at work. In doing so they were able, albeit briefly, to reclaim their time, autonomy and in some cases their dignity, despite the many attempts to monitor their activity.
Workers who are subject to algorithmic control have taken it upon themselves to exploit the organisation of their work to push back against their respective managerial regimes. It’s not that I think unionisation is a redundant consideration, but rather that other political forms – which are already being adopted in advance of mass union membership – are seldom examined politically as tactics in their own right.
To judge workplace resistance in terms of its capacity to transform society would be to place an impossible burden upon it; but we can still consider the significance of workplace resistance for those who care about the transformation of class society. As Harry Braverman reminds us, it is imperative that we do not simply accept ‘what the designers, owners, and managers of the machines tell us about them.’ This also extends to the ‘objectivist’ view of technologies of organisation taken by trade unions.
By extending the principle further, to the political actions of workers themselves, we can see the gap that exists between official efforts to improve conditions in the sector – which either occur outside the workplace, and neglect to focus on the conditions of working life, or take a wholly uncritical view of things like time and motion studies – and the actions being taken by many workers on a daily basis to maximise their interests in spite of efforts by management to mitigate the ‘problem’ of labour.
By neglecting everyday resistance and other forms of ‘organisational misbehaviour’, we neglect those forms of workplace conflict that do not fit into a rather narrow organising repertoire. Worse still, a failure to see resistance and misbehaviour at work – or diminishing its relevance – preserves the false perception that managers have ‘acquired effective techniques of behavioural control’. These are good reasons to reframe the idea of resistance to algorithmic management if we want to develop a political strategy around it. As the labour sociologists Stephen Ackroyd and Paul Thompson argue, the idea that there is now no alternative for workers but total compliance simply does not square with what we know of organisational history, nor with the empirical evidence we have. It also raises the question of what exactly we count as resistance at work.
There are limitations on the term ‘resistance’, evoking as it does a reactive or negative posture. I use it in the way the sociologist Randy Hodson outlines, to refer to acts that are ‘intended to mitigate claims by management on workers or to advance workers’ claims against management’. Such a definition allows us to shift our focus from a primarily negative stance of refusal towards a more positive conception of struggle. It allows us to think about resistance through the lens of subversion. Subversion means we can think about misbehaviour beyond behaviour, resistance beyond negation, and disruption beyond interruption. It allows us to consider action that might take any of these forms, but which might also be understood as an intervention or creative redirection intended to bring about new conditions and to maximise workers’ space within the organisation in such a way that they can advance their interests, even for a moment.
It is important to say that workplaces are politically messy places. Resistance does not conform to an ideal type, just as there is no ideal resister. Even the staunchest opponents of managerial regimes can be tangled up in contradictions, ironies and unintended outcomes, while employees often ‘consent, cope, and resist at different levels of consciousness at a single point in time’. When politicising resistance we must tread carefully, particularly when workers’ actions may not have been undertaken for explicitly ideological ends. As Jamie Woodcock notes, there have been attempts to reframe anything short of complete compliance as sabotage. But whereas sabotage is intended to disrupt crucial mechanisms or machinery, many of the actions we see ‘do not significantly undermine the process of capital accumulation’. Nonetheless, they may still be political – if, for example, they are aimed at reappropriating personal dignity robbed by managerial practices.
Less formalised political activity is harder to see, and is not always explicitly ideological. But, while such actions may have a range of motives, this doesn’t mean we cannot think about these practices in terms of their amenability to collective action. As Mario Tronti suggested, the breadth, possibility and radical contingency of working-class struggle can be understood as refusal. For autonomists, refusal marks ‘the beginning of liberatory politics’, because it implies the agency of workers within the class relation, particularly the wage-labour relation. In other words, refusal is a key potential arising from the indeterminacy of labour power, the fundamental problem of management.
The first radical aspect of refusal, argues Tronti, is that it marks a point of departure from capitalist logic. It is the point at which ‘the working class confronts its own labor as capital, as a hostile force, as an enemy’, marking not only a point of departure for class antagonism, but a starting point for ‘the organization of the antagonism’. Long before we can conceive of revolutionary vehicles for overturning the present state of things, Tronti sees workers’ disillusioned, passive non-compliance with work as the spontaneous first step in refusal – the point at which the worker first refuses to be an ‘active participant’ by ‘opting out of the game’. For Tronti, this is fertile ground. ‘Hence’, he writes,
What appears as integration of the working class in the system, by no means represents a renunciation of the struggle against capital: It indicates a refusal to develop and stabilize capital beyond certain given political limits, beyond a fixed defensive cordon, from which aggressive sallies can then be launched.
His idea is that the transition from workers’ diligent activity to alienated passivity at work is at the same time the beginning of active refusal.
— An edited excerpt from Cyberboss: The Rise of Algorithmic Management and the New Struggle for Control at Work by Craig Gent.
[book-strip index="1" style="buy"]
