Workers shouldn’t accept the imposition of AI in the workplace as natural or inevitable.
Interview by Marta Checa, former co-editor of Equal Times and a multi-media journalist.
Cross-posted from Equal Times
Jason Resnikoff, a US academic currently teaching contemporary history at the University of Groningen in The Netherlands, specialises in the intersections of labour history and technology. He is the author of Labor’s End: How the Promise of Automation Degraded Work, in which he exposes how automation intensified human labour and eroded workers’ power, recasting technological progress as both inevitable and apolitical.
In this conversation with Equal Times, the former United Auto Workers (UAW, USA) organiser reflects on the impact of artificial intelligence (AI) on the world of work, and the individual and collective responses it may require. We dive into what counts as progress and who defines it, as Resnikoff challenges us to think about the different ways in which we might organise our lives and work – if only we could be liberated from the endless loop of technological escalation that we currently live in.
In your writings, you draw a parallel between the ‘irruption’ of automation in the 20th century, and AI today, starting with the term itself. You say that its definition was “vague” then, but the perceived effects for working people were nevertheless real. Could you elaborate on that?
The origin of the term automation is primarily ideological and I would argue that what the whole discourse around automation did was to set up a model for how corporate America would talk about technological progress, and how it would use that narrative to its own advantage. In other words, automation provided a way for capital to harness this widespread faith in technology. It took something people already believed in – this old narrative, older than the 20th century, that there’s something called ‘technological progress’; and that technological progress and human progress are the same thing — and put it to work for employers.
And that’s what automation really was about: a way of talking about machines that made certain economic and political outcomes seem natural, even inevitable. Now, of course, there were real technological changes, and there were real effects for workers. But we shouldn’t fall into the trap of imagining that the machines themselves are the agents of change. They’re not, it’s the people in charge who make decisions.
And you see similarities with AI…
That’s why I think the current conversation about artificial intelligence sounds so familiar. The rhetoric around it is remarkably vague, and that vagueness is doing a lot of work. When you look at what AI researchers actually say, they’re usually very precise: AI isn’t a single technology, it’s a discipline, a way of thinking about computation. But when employers (and CEOs) talk about AI, it suddenly seems as if a technological revolution is taking place, and that’s a very different story.
Take Elon Musk, for example, when he says: “AI will do it all.” What does that even mean? If we look at the real technologies – large language models, machine learning, natural language processing – they are narrow tools. Useful, yes, but not revolutionary in the way the rhetoric suggests. And that’s the key point: this vagueness (of the term AI) allows powerful actors to control the story of technological progress, and lets capital keep control over who gets to benefit from it. That’s exactly what the automation discourse did in the 20th century, and it’s exactly what the AI discourse is doing now.
Sure, some jobs will change. Some might disappear. That’s always been true under industrial capitalism. In the end, the technology changes, but the logic doesn’t.
How do we, in a way, unmask this? Was it perhaps easier to see through before – as you could see the machines doing the work, compared to the complexity of what involves AI?
I actually don’t think it was clear in the middle of the 20th century what these machines were really doing. And that uncertainty was part of the problem. You could walk into a factory in Cleveland in the 1950s, and it might look like the machines were doing most of the work. There were still thousands of people working there, but what you couldn’t see were the jobs that had already been relocated – first to the South, and later overseas. And even within the factory, it might appear that there were fewer people working, but what you wouldn’t notice was that the guy working down the line was now working harder for less pay. One of the defining features of these ‘labour-saving’ machines is precisely that they obscure labour. Now, that’s not to say that machines can’t replace labour – they do. But employers also use technology in other ways: to outsource work, to reorganise it, and to make people work harder.
So “how do we unmask this?” – well, it needed to be unmasked then, too. This has always been part of the story of technology under capitalism: new machines arrive, and with them comes confusion about what’s actually changing. It happens that workers can sometimes be working harder, and still believe that it is automation that is doing the work. When desktop computers started showing up in offices in the 1980s, workers suddenly found themselves doing administrative work they’d never had to do before. The technology creates this aura that makes the extra labour feel invisible – you don’t necessarily or immediately sense that you are doing more work – but in practice, you are.
So, how do you cut through all that noise and make sense of what’s happening?
When it comes to technology today, there isn’t much solid analysis coming from the left – including unions – for working people to pick up. So, unsurprisingly[TG2.1], when a company like Apple or a figure like Elon Musk introduces a new gadget or system, people tend to respond with awe rather than suspicion.
The story we’ve been told for generations is that progress comes from capital, that every new machine is a sign of civilization advancing. And workers, for the most part, have internalised that story. What I think we need – and this is especially true for unions and educators – is a way to separate those ideas. We need to make it clear that technological progress does not have to come from capital. It’s not their exclusive domain. And just as important, we don’t have to accept their definition of what counts as progress.
We also need to ask: what does progress actually look like for you? Because if your idea of progress is a robot slave – that’s effectively what’s being sold to us – then maybe that’s worth rethinking. But I doubt most people would actually say “yes” to that. Take Elon Musk and his Optimus project. Why should his vision of progress be the one we all have to live with?
We could agree that there are certain jobs, and a few tasks within many jobs, that we would rather not perform, and there we are happy with technology coming our way, regardless of its origin.
There are aspects of many jobs that people don’t enjoy, even if they still like their work overall. Another way to put this question is: there are necessities – tasks – that people simply don’t want to do. Some jobs, historically, have been extremely difficult to make humane. Take mining: historically, only coerced laborers, very poor people or people without other options became miners. It’s hard to imagine that kind of job being made ‘good’. So of course, there are cases where a machine might be better suited to do a task.
I’m not a Luddite, nor anti-technology. If a machine can make a job safer or more humane, that’s obviously a good thing. But the deeper question is: who is designing these jobs in the first place? People decide socially what is “necessary”. There’s a term for that: social necessity. Many of the jobs, as they exist, are socially determined. Think about housework – or ‘social reproductive labour’ – the kind of work that’s been historically devalued and feminised. Today, we have all sorts of machines that are supposed to make domestic labour easier. But, as the historian Ruth Schwartz Cowan showed in her book More Work for Mother, these devices often create new kinds of labour instead of eliminating it.
At the turn of the 20th century, a group of thinkers asked: ‘Why are we each doing all this domestic work individually?’ They pointed out that industrial capitalism had already centralised the production of cloth. So why couldn’t we apply that same collective model to other tasks of social reproduction? And then they imagined collective childcare, communal spaces that could distribute reproductive labour – designing a better system of care. That’s a crucial point. When the people who hold power design the work process, their model almost always includes hierarchy and compulsion. So the machines they build – the technologies they promote – reflect that same logic. But it doesn’t have to be that way. We could imagine machines and technologies designed around different principles.
In today’s reality, how could we possibly change the course and have this reflection?
We’re indeed operating in an extremely dynamic environment. It is also one where Big Tech has completely flooded the zone. And all of this is happening at a time when a lot of people already feel pretty powerless. In this environment, I would propose an exercise, focused on workers: let’s just imagine for a moment that there was no more technological change. It stops today and nothing new can arrive. “How would I organise my life, my job? How would I organise my politics in a way that I thought was just and good? What would I need?” And you might end up saying I would need to invent a machine that can do X or Z. But I imagine most people would say other things, like “I’d want to control my time”.
But you have to begin by having your own idea of what it is you want to achieve. And the reason all these apps and things make it hard is because they have premises in them: “Don’t you want to achieve or do this?” They’re telling you what you might want to do, and they’re also trying to shape what you might want. Each tool carries its own agenda. Maybe it’s about making money, collecting data, or locking you into an ecosystem of software. Maybe it’s just a developer trying to get their app noticed. But in any case, those motivations are not necessarily aligned with what you, as a worker or as a person, actually value.
This exercise returns the idea of the political to our working lives. And for me, as a labour historian in particular, what I kind of find impressive time and time again is how unpolitical our working environments often appear.
Technological changes are not neutral or apolitical, you mean.
No, indeed. Workers and managers end up navigating the same environment – one defined by the constant pressure to adapt and to “not fall behind.” Even managers who don’t particularly like new tech feel that pressure. They might not initiate the changes, but they’re still judged by them. And they, in turn, will judge their workers according to how quickly they adapt. That’s how technology becomes political. And behind all of that, Big Tech itself is actively shaping the working environment to its own ends. They have this hegemonic power over our built environment, then, using their apps becomes a kind of collaboration.
But again, how could any of us challenge anything?
This isn’t an individual problem – it’s a collective one. The way people interact with the means of production has always been a social issue, not a personal one. If we want to imagine ways of engaging with the means of production that aren’t alienating, that’s not something you can do alone. That’s a huge project and it would necessarily need to be collective because you can’t disconnect from the industrial apparatus. You have to live within its logic, and that makes it incredibly hard to resist or reshape.
I’m not saying that individual workers can or should try to refuse that reality. What I am saying is that rather it should be a goal of organised working people to say we should negotiate over this issue. It should be back on the menu the way a lot of unions deal with it now. Because right now, most unions treat technological change as something that just happens. But if unions don’t negotiate over these changes, and if the state doesn’t step in to plan or regulate them, who will?
Elaborating about trade unions, how do you see their role in this technological and power shift?
Right now, a lot of unions are thinking seriously about AI – whatever that is – and trying to figure out a good policy on it. But that’s a little different from saying: “The union should set up as a priority having some say over how the work is done in a meaningful sense”, or “We want to have the power to veto new technology”. I think it’s been a mistake on the trade union movement’s part, to abandon the idea that workers should one day control the means of production. They not only gave up the fight – because that was not a fight they could win – but worse, they also gave up the analysis that came with that objective. They began to agree with their bosses that their bosses were the ones who would produce a better technological world. That’s a huge concession.
And now it is much harder for unions to respond to technological change, because they don’t have a clear goal when it comes to the built environment in which people are working. Except “we want to protect the jobs”. But that’s a very basic thing. The big problem from the beginning of 1950s is that labour leaders were afraid that they would be called “anti-progress”. To reject a new technology would be the same thing as saying: “We want to reject economic growth”. But that fear came at a cost. Because if labour leaders had developed their own idea of what economic growth meant, the neoliberal turn would have been much harder to sell.
So, I think that’s where unions need to change. Power might come later, but clarity of purpose has to come first. Imagine if you’re a young worker today, and you’re wondering whether to join a union. What if the union said to you: “Your workplace is alienating, and we want to make sure that one day you can control your own life, and that includes controlling your environment”. And compare that to: “We want to make sure that you can enter a retraining program if they take your job.” It’s a much less inspiring talk.
How different is the challenge for them today though?
I’m not sure this challenge is unique to the current technological moment we’re in. It feels more like an extension of an older problem – one that really took shape in the 1970s and 1980s, with globalisation, when many states agreed that capital and goods will move freely across borders but workers will not, nor will workers’ rights.
Technology has made it possible for production to be globally distributed in real time. The ‘workshop’ is no longer contained within a single country or continent. You can have people working on the same labour process, scattered across the world, and that’s extremely difficult to organise.
How then should unions react to this specific situation?
One strategy is trying to make the labour process entirely local – to one specific place. The other way is organising along the supply chain. That’s much harder, but it’s probably the only realistic way forward in a world where work is networked rather than centralised. A good example is the movement to organise Amazon; it isn’t a single factory but a logistics network of warehouses, delivery hubs, data centres and subcontractors that stretch across borders. You can’t organise that company plant by plant; the structure itself is designed to defeat that strategy. You have to map the labour process and organise horizontally along its many routes.
And we’ve already seen what happens when traditional methods are used. When workers organised a warehouse in Quebec, Amazon’s response was simply to shut it down. They’d rather abandon an entire region than deal with a union. They’re that powerful that they can go around it. And I suspect the same will be true when it comes to tech. We’ll need to map out the labour process – where data is generated, processed, and where profits are extracted.
Can you think of any successful example of a trade union fight that has effectively met the challenges coming with AI?
A couple of years ago, when the Writers Guild of America went on strike, one of their major bargaining points was that they did not want AI writing scripts. They drew a clear line and said: “This is our work. This is what makes it meaningful.” I thought it was a courageous position to take. And it didn’t make them look like Luddites, nor irresponsible. If anything, it was understood that they were protecting a craft. It also helped that, at that moment, AI still did not look inevitable. And importantly, their position looked all the stronger because of who they were negotiating against. The studios are not exactly sympathetic figures, their motivation was obviously profit.
And even though they got a half measure, the fact that they made that demand at all was significant. They showed that it’s possible for a union to challenge the technological narrative directly, to say ‘no’ without looking anti-progress. I think people still have some capacity to decide what their own values are, and to make arguments for things that should be valuable. And I don’t think we should surrender that capacity to what capital thinks is valuable. The labour movement has a responsibility to, at the very least, be able to say what it actually wants. They might not get everything they want. But the point is to articulate it.
Be the first to comment