![Craig Gent, Cyberboss: The Rise of Algorithmic Management and the New Struggle for Control at Work (Verso 2024), 256pp.](https://uploads.counterfire.org/uploads/2024/08/cyberboss-lg.jpg)
The important subject of management’s use of technology to increase control and worker resistance needs a stronger conceptual approach than that in Cyberboss, argues Kevin Crane
For many, many years, I used to be baffled as to why so much of the left seemed actively disinterested by technology. That turned around a little under a decade ago, and it is now common to see analyses of technology from explicitly critical and political perspectives, which really can’t be a bad thing. Cyberboss is simply one of the latest in what is now a fairly extensive genre of books taking a look at the role of modern electronics and software in people’s working lives. Its subtitle is The rise of Algorithmic Management and the New Struggle for Control at Work which is a relatively specific theme, and one that other writers have tackled. That wouldn’t be a problem in itself, except that I don’t think it really encapsulates either what this book tries to do, or what it actually does.
Workers’ inquiry
The early parts of the book lean heavily on its strongest elements: the accounts of people working in modern, high-technology workplaces. Author Craig Gent is using a version of the ‘workers’ enquiry’ method of workplace journalism – getting stories straight from the ground-level of industry from people with the most basic roles – and makes a decent go of acquiring some good anecdotes. You get a strong flavour of what working for a delivery app or in a logistics warehouse is genuinely like on a day-by-day basis, and how the various tracking and dispatching methods impact the human beings subjected to them.
Anyone who’s been following the struggles for unionisation at companies like Uber and Amazon will have an idea of some of the petty tyrannies and injustices built directly into the management technology. Productivity targets are often intentionally obscure and frequently practically impossible, to the point where it is questionable if they are really ‘targets’ in a true sense, since they do not seem to be intended to be actually met. Gent lightens the mood of these with some fun stories of resistance, and plain old mischief. We get stories about workers developing a wide range of means for alleviating stress and exhaustion through subtle fiddles. One particularly memorable power play by one of the warehouse workers is intentionally screwing up every single DVD order he gets by replacing whatever they’ve asked for with a copy of Rogue One: A Star Wars Tale. I can see how this would be quite a funny prank, dragged out over a sufficient length of time, though I do question what its radical potential is in any meaningful sense.
These accounts don’t ultimately feel like they add up to that much, however. Gent is particularly interested in things that don’t feel very relevant to me, such as using technical glitches as opportunities for ‘workers’ resistance’, which he justifies by pointing out that all technology will have faults, so it makes sense to use those moments to express self-determining power. I agree that technology will never be fully free of malfunction, but that’s not a novel feature of modern high technology (it was just as true of the spinning jenny), and you can’t spend all day every day waiting for a glitch in the matrix to spur you into rebellion.
There are also slivers of accounts I’m not so keen on. This isn’t a book specifically about trade-union activists – fine – but that made me bristle at one particular story. We hear that a group of warehouse workers attempted to push down their productivity as a protest against multiple grievances, and the author casually drops this line in:
‘The action was conceived by non-union salts (politicised workers who join a workforce in order to organise it) but was joined by an initial group of supportive workers, before spreading to include most of the agency workers on-shift’ (pp.192-3).
I beg your pardon? ‘Salts’ is an obscure American term, and is not used in the entire book at any other point, leaving a swirling storm of questions in my mind. The first thing that occurs to me is that this sounds like what socialists sometimes call ‘substitutionism’, literally getting activists to run into a space and do things on behalf of workers rather than them doing it themselves. Further, from the perspective of the book’s analysis, I think this throws the study that is being done into question. Now, I’m an engineer not a sociologist, but I do think if you’re analysing a system, external factors like that need to be considered as potentially distorting of the results you think you’re seeing. Actually, while we’re on the topic of trying to speak to things you aren’t expert on…
Technical limitations
The author runs into trouble very early on with his attempts to discuss the actual technology, and I get the sense he knew this, but had gotten too deep in his concept of ‘algorithmic management’ to give himself room to be able to write himself out of the bind. Further, I think he fell afoul of the very current mystification of the word ‘algorithm’, which is a trap you can exit very easily by just substituting ‘the process’ or ‘the bloke in charge’ for it in your head.
Ascribing godlike qualities to these systems obscures all and reveals nothing. A strong example is the way that the means by which taxi and courier apps exploit drivers being described as if it were a triumph of seven-dimensional gamification … when it’s actually just done by keeping way too many drivers on the roster for actual customer demand.
We get passages fetishising ‘the algorithm’ into the role of the actual manager at points, only for these to be back-pedalled later because all these employing companies do have people employed in management grades. Them communicating mostly via text message rather than face-to-face doesn’t functionally change their role that much. Gent has to concede at various points that the features he’s ascribing to ‘algorithmic work’– such as being told not to talk – are not unique to it, and this becomes a slightly repetitive handwave. The real problem this concept has is that the things he wants to call ‘algorithmic work’ cannot be specifically defined by the existence of an algorithm… because all work always is and always was done according to an algorithm (an algorithm boils down to being instructions for a series of steps, or a planned process). Algorithms are not a new thing brought about by computer technology; computers are a technology that were invented to help do already very necessary algorithms better!
Attempting to sound technical – without really getting it that much – leads to some seriously ill-advised passages, such as here when referring to some low-level disruption by workers:
‘These moments of accidental resistance reflect resistance as we would expect to find it within an electrical circuit: they hold up the flow of work and – in the case of technological disruption – information that is essential to the algorithmically managed workplace’ (p.188).
Electrical resistance is a spectacularly bad analogy for human self-determination, for the simple reason that resistors are undynamic. The resistance in an electrical circuit is what guarantees that at least some energy will be dissipated as hot air as the current flows. The quantity of hot air will be a fixed value proportionate to the current. The aspects of a circuit that force it to change state over time are capacitance and inductance. If, as I presume you would in the context of this workplace metaphor, what you really want is for a circuit to malfunction catastrophically, you should actually take the resistors out, since they are providing crucial stability!
Ironically, I think this misguided metaphor does fit certain aspects of the theoretical framework of the book to an extent. It’s just it does in the opposite of a way that supports the arguments being made.
Autonomism’s legacy of liberatory confusion
Quite early on in the book, Gent starts bringing in aspects of Marxist theory, but not of a good kind. I’m not saying he picked exactly the worst choice of political tradition possible, but it’s definitely the most baffling. He has decided to dust-off autonomism to provide the intellectual spine for Cyberboss, a good decade after virtually everyone had stopped talking about it, and if nothing else, it serves as a reminder why they all stopped.
Autonomist Marxism is a formerly fashionable set of theories that emerged from Italy in the wake of a series of defeats of that country’s once-powerful Communist trade-union movement, and really took off after the 1990s. As post-Cold War globalisation meant that established workers’ organisation was foundering, a wing of the left intelligentsia started to argue that the old point-of-production was not the focus of class struggle, which was now simultaneously everywhere and nowhere. Everything you did under capitalism was now work… or/and resistance to work. Strategy became a dirty word to autonomists, who increasingly came to believe that denying the ubiquity of struggle anywhere would hold it back everywhere.
Aware that few of his target audience will have even heard of autonomism by this point, Gent struggles to introduce it to them. His problem is not the same as his issue with algorithms; those can be understood. The autonomist concept of operasti has never been successfully translated from Italian to English because, well, it is really more of a vibe than anything on which you can actually base an organisational method. Gent’s attempts to build on this ideological vapourware result in some flourishes of astonishing purple prose, that will be very familiar to anyone who remembers any autonomist writing from the days of G8 protests.
None of this is helped by Gent getting some aspects of very recent history dead wrong. He has autonomist theory and digital network activism peaking in popularity at the same time, as a result of the 2008 economic crash. In fact, autonomism began to decline in reach after that period, initially because of the defeats of the ‘horizontalist’ movements that sprang up in its immediate wake and then even more so as globalisation began to visibly decline, and the left took turns toward populist electoral organisations. Conversely, the concept of networked digital organising didn’t really catch on for a few years after 2008, because Web 2.0 social media was still quite new, and smartphones were not ubiquitous until some years later. It might seem logical that the autonomist movement could have pioneered online activism, but that’s not what happened because material circumstances didn’t play out in that order.
Detours to a dead-end
If Gent’s resistance metaphor does anything, it is somewhat unwittingly reproducing the old autonomist hobgoblin of workers’ struggle being everything and nothing, and in that sense, it would just be a stable feature of the overall system like a resistor in a circuit. He gives a name to the form of struggle he believes he’s identified – ‘Metic resistance’ – and tries to argue that it is both a novel phenomenon that is specific to ‘algorithmically controlled work’ and something that the trade-union movement has yet to pick up on and realise the potential of. Right at the epilogue we get this line:
‘…as I have argued at the beginning of this book, both transparency and health and safety are thin ledges upon which to hang our concerns’ (p.215).
If campaigning on health and safety at work is a ‘thin ledge’, what does that make the practical joke of swapping out DVDs for a naff science-fiction film? At least that’s mildly amusing – elsewhere we get told that simply doing jobs as badly as possible is a form of resistance, and Gent chides trade unions for failing to endorse it. But that would never work: my own experience is that where individual workers do manage to use intentional negligence as a power move, it’s as much a tactic directed against co-workers as it is against management. In any case, trade unions must, by definition, reach compromises with management in return for the members obtaining gains: one of the compromises will be that managers do get the standard of work for which they’ve asked.
I actually do appreciate that quite a lot of work, particularly research, went into Cyberboss, but I simply cannot find a level on which to recommend reading it. The interview material with actual workers has some insights, but honestly could have been distilled down to article length without losing much. Despite the book’s short length, individual chapters manage to be paradoxically far too long and lurch off at baffling narrative tangents: the fourth chapter is titled ‘Technological Politics’, and yet concludes by telling us we’ve got to read more research before we can get to the politics. I suspect some people may pick up this book expecting to get a technical explainer, and they really won’t find one.
The biggest problem, though, is just trying to bring back autonomism: it really didn’t deliver on anything it promised a quarter century ago, and that’s when some of its core conceits were surface-level plausible. I really don’t know what would attract anyone to this in 2024.
Something I found really telling was the book’s epilogue: suddenly the topic of conversation is the struggle of the Hollywood writers’ union in last years’ massive strike, which was in no small part against the encroachment of generative AI on writers’ work. Now, this was a huge deal, and worth talking about, but there are two problems. The lesser of these is that the book had empathically not been about AI up to that point. The greater problem is that it was a very conventional mass strike, by a very traditionally and densely unionised white-collar workforce that in no way resembles the mostly unorganised blue-collar workers that the book is actually about, and consequently in no way serves the arguments in the book. Recent breakthroughs in unionising Amazon and Uber workers are conspicuously not present in this epilogue, and I really suspect that that’s because these would have served the intended narrative even less. It really cannot speak well of any strategic or tactical conclusions that this book might promote to close it out with a such an avoidant non-sequitur.
Before you go
Counterfire is growing faster than ever before
We need to raise £20,000 as we are having to expand operations. We are moving to a bigger, better central office, upping our print run and distribution, buying a new printer, new computers and employing more staff.