Big Tech is selling us a future in which workers are increasingly disposable, human expertise is devalued, and corporate power is more concentrated than ever. But is that future inevitable? In this episode of Movement Memos, I talk with Alex Hanna, co-author of The AI Con, about AI as a tool of class warfare, the hype that obscures its real-world harms, and how workers are organizing to resist Big Tech’s agenda and win meaningful protections in the workplace.
Music: Son Monarcas and Guustavv
TRANSCRIPT
Note: This transcript was originally published in Truthout. It is shared here with permission.
Kelly Hayes: Welcome to “Movement Memos,” a Truthout podcast about organizing, solidarity, and the work of making change. I’m your host, writer and organizer Kelly Hayes. Today we are talking about how workers can resist automation and AI in the workplace. We’ll be hearing from Alex Hanna, co-author of the book, The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want. Alex is a writer and a sociologist of technology, labor, and politics. Alex is also the director of research at the Distributed AI Research Institute, and leads the organization’s Luddite Lab project. The greatest promise of AI in the business world is the promise that it will eliminate workers. Some people have already felt those impacts. But far more often, we are seeing AI leveraged to surveil, control, and devalue workers. So, how is AI impacting our work lives, what’s on the horizon in the struggle, and how are some workers fighting back and winning?
We’ll get into that and more as we discuss what we’re up against, and what kind of political education workers need to wage a strong fight.
If you appreciate this podcast, and you would like to support “Movement Memos,” you can subscribe to Truthout’s newsletter or make a donation at truthout.org. You can also support the show by subscribing to “Movement Memos” on Apple or Spotify, or wherever you get your podcasts, or by leaving a positive review on those platforms. Sharing episodes on social media is also a huge help.
Truthout is an independent news organization, publishing stories that the craven corporate press won’t touch. We are a union shop with the best family and sick leave policies in the industry, and we could not do this work without the support of readers and listeners like you. So thank you for believing in us and for all that you do. And with that, I hope you enjoy the show.
[musical interlude]
Kelly Hayes: Alex Hanna, welcome to Movement Memos.
Alex Hanna: Thanks for having me, Kelly.
Kelly Hayes: How are you doing today?
Alex Hanna: I am recovering from whatever is in the air, but I am soldiering through.
Kelly Hayes: Well, thank you for joining us while you’re still recovering. I really appreciate it, and I’m so glad you’re on the mend. I know a lot of our listeners will be familiar with your work, but for the unacquainted, what would you like people to know about who you are and what you do?
Alex Hanna: Yeah, so I am a sociologist and writer on labor, tech, and politics. I’m director of research at the Distributed AI Research Institute, and we’re a research institute that researches the present-day harms of AI. I feel like I have to say that now, because there’s a lot of imaginary harms that are in the conversation now. And we also develop community-based technology. And so then what I focus quite a lot on is understanding the connection of AI and labor and work, as well as working and understanding the kind of ecosystem of who’s in the AI space and what political incentives that they have. And so I have a book called The AI Con, which I co-authored with Dr. Emily M. Bender. And I also have a podcast with her called “Mystery AI Hype Theater 3000.”
Kelly Hayes: Which is such a great name for a podcast. I am really excited to talk about your work, and in particular your work around AI and labor, including the work you’re doing with the Luddite Lab. But before we go deep on labor, I feel like we have to wade through some of the AI headline muck that folks are dealing with right now.
Recently, Anthropic co-founder Jack Clark declared that “the singularity” has begun. He made this announcement at a gathering of economists and policy makers at the IMF [International Monetary Fund] and World Bank meetings. And earlier this month, Anthropic announced that it was prohibiting any sustained cruelty towards its AI models amid an ongoing debate within the industry about whether these systems might be conscious or have some kind of moral status. (Spoiler alert: They do not.)
So we’re being told on the one hand that these technologies are ushering in a transformation of human civilization that’s beyond our control or soon will be. And on the other hand, we’re being asked to contemplate whether the products themselves might be sentient beings whose welfare we need to consider. These extraordinary claims, which always seem to come from the people building and selling these products, are often received with great seriousness by the press, which I think only compounds public confusion. So, what are your thoughts on this?
Alex Hanna: Yeah, I mean, I have a lot of thoughts on this. I mean, I have thoughts on the claims of singularity or AGI, which is artificial general intelligence. I have thoughts on Jack Clark as a person and a figure in this space. And then thoughts on the ways in which they’re tied up together. Jack Clark is kind of an interesting figure because he used to be a policy person at OpenAI. And then I actually have been reading [or have at least been attentive to] his newsletter for years now. And it’s kind of interesting the ways in which he has gone from someone who had some interesting things to say about AI and interesting things to say about policy, to someone who’s very much bought into the narrative of things like the singularity, this hypothetical event in which human beings all merge into a singular kind of consciousness with the machine.
This is an idea that’s been proffered by someone named Ray Kurzweil, who is a futurist who has worked at Google for the last number of years. And it’s this fantastical idea that comes from science fiction. And what’s not new is that there’s claims about the singularity is already here, the same thing with artificial general intelligence. Oh, it’s already here. This is something that [OpenAI CEO] Sam Altman has said. And so you have this thing where there is this model of the world that at one point, is eschatological, like it’s: We’re going to have the rapture, we’re going to have the end times. And now as an economic move, these people within these companies have said, “Well, it’s already happening. We’re starting to be raptured. We’re starting to be moving into this space.”
And you’re like, “Ah, okay.” So your model of the world has changed from kind of an end times type of Christian, everybody’s going to merge into this one thing and there’ll be this tipping point to, well, it’s actually already happening and this is very important. And it just gives the game away that Jack Clark is giving the speech to the IMF, this neoliberal organization that has done so much damage in the majority of the world through structural adjustment programs and the ways that they are forcing austerity spending by their client governments. It’s effectively saying we need to revamp this element of trade policy or whatever the IMF is funding to accommodate for this kind of fantastical thinking. And so that’s definitely one thing that that signals.
The element of the model welfare thing, the idea that the models could just feel pain, I mean that’s this kind of thing which is — this has been around for quite a bit of time. And there was this thing that happened a few weeks ago where there was this GitHub repository … So GitHub is a code-hosting platform that is currently owned by Microsoft. And there was these sets of experiments that were created when this person who read this paper that said, “Well, if we were going to simulate pain in a language model, this is what it would look like.” And so this person developed this repository, this bunch of code where they were simulating intense amounts of pain that would be afflicted to a model. And then it was just an incredibly ridiculous event because there were many people — many users on the platform were then targeting this person saying, This person is literally torturing models and they’re putting the code on the web. And what an inane timeline because now there’s this focus on this idea that models are aware and they’re conscious and they can feel pain.
And to what degree are we — what evidence is given that a model can do that? These are big piles of linear algebra, that is effectively, I mean, that is not what is happening. And what really gives it away to me too is these are people who are more defensive of big piles of linear algebra than they are of real human beings, of Black and Brown people, of people in poverty. And this has been said by a lot of people in many different guises, but it’s effectively like people in this field seem to care a lot more about hypothetical, maybe conscious piles of linear algebra, rather than people in actual pain in the world right now. And that just tells you all you need to know about their ideology and how they view the humanity of people who are in the margins.
Kelly Hayes: Absolutely. And I think all of this goes back to how we have to think about AI conceptually, because we’re being pulled in a lot of imaginary directions. So it’s important to stay really grounded. You’ve argued in your work that AI isn’t even a coherent set of products and services; it’s a marketing term. It’s used to describe not just chatbots, but a wide range of computational services, including things that existed well before the era of chatbot hype. And AI, in addition to being a sort of product genre, should also be understood as a political project. It’s at the center of an effort to reshape social and economic relations and create new forms of dependency and control.
The devaluation of human beings, our thinking, our well-being, and our labor is the crux of that project. And what you said about how we’re being told to fret about the emotional well-being of models, when actual living, breathing people in the margins don’t get that level of concern from these tech oligarchs — it really is emblematic of what’s happening in a larger sense. We have hyper-exploited people in the Global South, whose labor keeps these platforms running, and never once have I heard one of these tech leaders say: We have to stop and think about how our use of these platforms is harming these human beings, because they’re suffering. Just like you don’t hear them talking about the people with asthma and other health problems who are negatively impacted by data center pollution. This project wants us to forget those people, to forget each other, while centering the well-being of the technology. We’re meant to see ourselves and each other as less relevant, and less important to the future.
So there’s a way in which the cults of Silicon Valley are feeding their visions and aspirations into a more digestible popular discourse. You have effective altruists and accelerationists and people who see humanity as a kind of bootloader species, for a future where space is populated by trillions of digital beings. And this weird, fashy, sci-fi fantasy really adds up to the devaluation of most living people, here and now. And so, in some ways, that’s the most important thing to understand about any aspect of AI as a marketing effort, and as a political project, is really that it’s about our devaluation. That’s what every one of these talking points comes down to: how much AI is worth, even if it’s not turning a profit, and how little each of us matters, in the new scheme of things.
Alex Hanna: Yeah, absolutely. I mean, I think that’s really hitting it on the nose. If you envision a few different elements of AI as a political project, you can see this devaluation of the human in a lot of different incarnations. And so I focus a lot on work and understanding that the primary economic rationale behind AI is obviating the need for labor. If there is some kind of way in which you can get rid of or automate so much of the work of human beings, then that means that you can just do massive layoffs and you don’t have to have folks who are real skilled laborers involved in the work process. And we’re seeing this in so many different industries, whether that’s health care, whether that’s education, whether that’s software engineering, whether that’s public service, whether that’s in social welfare. I mean, the list kind of goes on and on.
And there’s so many people who are salivating at this, especially venture capitalists who are willing to fund these projects to the tune of millions of dollars in their first funding rounds. And so we see this in labor. We see this as people who are living in the shadow of data centers. We see so much of the contention — not only in the U.S. but around the world — about how everybody needs to kowtow to the “greater good” of data center development in the U.S. We’ve seen it very sharply in the fight in Memphis, in Boxtown, which is a majority-Black neighborhood that is experiencing just a renewed round of environmental racism based on the build-out of the xAI data centers. We’re seeing this fight in the Philippines. We’re under this program, which Jacob Helberg [Under Secretary of State for Economic Affairs] has headed this program called Pax Silica, which is intended to be this anti-China AI supply-chain securitization scheme.
And one key piece of that is the Philippines, where they want to build more data centers and semiconductor manufacturers. And this is against the wills of Indigenous Filipinos and land defenders there. And so they’re fighting with their own government and also this Pax Silica arrangement or alliance that wants to effectively ensure U.S. dominance against China. And so that’s being lorded again over people’s heads to say, “Well, we need this to defend X, Y, and Z.”
And so it’s this whole arrangement and it’s causing this reconfiguration of who is fighting who. And so now you have — elections are coming quite soon in which data centers, everyone’s opposed to data centers politically, from left to right, in that the elites are responding and they have to use some pretty old tricks. So referring to national security has been the way that people on the right have been trying to convince their constituencies.
And then Democrats are approaching this from a much more mealy-mouthed sort of, “Well, we want these data centers, but we need them to pay their fair share or ensure that there’s sufficient community benefit agreements.” But for Democrats, just abolition, just not having these data centers is not on the table. And so it’s wholly this political project on reconfiguring the centralization of political and economic power … [and] most people are stridently against this.
Kelly Hayes: It feels like a kind of full-spectrum class warfare, as a political project. Like there’s an effort to make existing arrangements of work and social life impossible to reproduce, such that we see a massive reorganization of society and a corresponding disempowerment of workers. It’s an effort to take what tech has done to the transportation market through rideshares, or to the retail market through Amazon, and apply that kind of disruption to entire economies and nation-states. Existing ways of operating get choked out, and new terms can be set by whoever controls the platforms and products that we’re all forced to depend on. Does that characterization resonate with you? And what’s familiar about this moment, in terms of how capitalism has always operated and reinvented itself, and what’s threatening in a new or different way?
Alex Hanna: Yeah. I mean, I think that seems like that’s what the desire is. I mean, we’re in a space now where the kind of class arrangements are very much set. And I mean, the kind of political realignment that’s happening is very interesting to witness in real time, where so much of capitalist desire is becoming made just abundantly clear in the ways that AI is leveraged as a cudgel against the working class and a cudgel against really anybody that’s holding land as well. I mean, that’s not just people who are land defenders that we see for spaces that have been traditionally held by the rightful Indigenous owners, or rather caretakers, but also the ways in which it’s also kind of a target of people in the U.S. who own lands or who are more conservative or farmers. And so I don’t have a good theorization of it to put a point on it, but you can very easily see how it’s just kind of a full-scale assault on anybody who’s not fulfilling the AI project of just kind of intense centralization.
And so what that means is these coalitions they’re developing are really novel and a little bit tricky, but some have been just very successful. And I mean the kind of rearrangements we’ve seen in places like New York State, where it’s this kind of combination of environmental groups and land defenders and First Nations and labor and democratic socialists have constructed in such a way that they’ve passed a moratorium in New York State for at least, I think it’s a year-long one.
And we’re seeing more and more of that tension coming to the fore as so many different groups are being opposed to data center build-out. And so yeah, I mean, we’re at this moment where AI is this manifestation of so much of the tech oligarchy and tech fascism where I’ve been in this area for 10 years and it’s just kind of an intense kind of thing seeing that, Oh, this thing I’ve been studying for 10 (I mean actually probably closer to 15) years is just now coming to hit us in the face by really presenting how much this has become a full-throated tool or full-throated incarnation of class warfare.
Kelly Hayes: Yes, it’s absolutely a class war offensive, and I think it’s so important to discuss it in those terms, whether it’s being wrapped up in these eschatological predictions about the end of the world, or talk of abundance, or the singularity. Whatever fantasies or talking points they’re throwing at us, that’s all wrapping paper for a project where capitalism is sort of redistilled — or, as Ruth Wilson Gilmore would say, the next round of capitalism saving capitalism from capitalism. In an era of catastrophic change, AI is being positioned as the basis for a reorganization of social and economic relations, using particular platforms and products to make human beings easier to isolate, control, disorient, and swap in and out of various labor roles.
So, we know that in the business world, there’s been a lot of excitement at the top about AI products because of the potential to eliminate jobs. But job elimination isn’t the only way workers are being affected. Can you talk about the changes you’re seeing, both in terms of job losses, and how people’s working conditions are being affected?
Alex Hanna: So we see jobs being eliminated because of at least the purported promise of AI tools in a lot of the creative classes. So we have seen a bit of a decimation of stable jobs in things like freelance graphic design, concept art. We see it a lot in translation as there’s been more of a move to use automated tools to do translation, even though these tools are not actually approximating good translations of anything. So we’ve seen, and especially in very high stakes, high-risk situations like in asylum hearings or in immigration. So we definitely have been seeing more of an encryptment of these tools there. We’ve seen it in software engineering and coding, where so much of the industries and the corporations who are actually producing and making these tools or supporting the data center build-out, they are doing massive layoffs of their own employees.
And we’ve seen this at places like Meta and Oracle. I think Oracle has been one of the places, and this is an organization that is supporting so much of the data center build-out, and they’ve just absolutely thinned out their senior developer pool. And I have some friends at Oracle who have mentioned that and they’re doing it in a ruthless way, people who are actually somewhat close to retirement or having retirement savings, actually just having those wiped out. And so yeah, I mean, those places are very, very worrisome. I mean, these are also places which have some of the worst union density, either because most of the fields are freelance or they work on a bit of a guild system. And so you’re only in the guild during the duration of a job.
And so we’ve seen a lot of that. And mind you, it is not necessary for the tools to actually work well for jobs to be decimated there. They effectively have to be thought to work well, and that may be just good enough for corporate managers there. They can then hire or rehire the same people at a fraction of the cost to fix those outputs, the sloppy AI outputs. And so we’ve seen another bit of a cottage industry even in fixing the outputs of those tools.
Kelly Hayes: I really appreciate what you were saying about turning people into fixers. It reminds me of what I’ve been hearing from artists, about how graphic designers are increasingly being sent AI-generated slop by clients who basically say, “This is what I want.” Then they expect the designer to turn the slop into something slightly better and a little more human.
So it feels like so much of this isn’t about taking human beings out of the workplace altogether. It’s about making our expertise an add-on, and making us more interchangeable. And again, so much of what actually gets discussed around this, in popular discourse, is just the packaging. It’s the sci-fi narratives, the fantasies, the promises, or the threats. Are the bots conscious? No, they’re piles of code. That’s a distraction. Are bots going to make us all rich? Of course not. There is no real agenda for “abundance.” That’s a fantasy pacification. Are the bots going to end the world? No, but people might, and some of our bad decisions could be outsourced to bots.
And lately, it seems like some of these tech leaders have belatedly noticed that all their apocalypse talk maybe wasn’t helping their cause, so suddenly we’re hearing more about abundance and universal basic income and all this shit. Meanwhile, they’re trying to create a tech mogul’s paradise, which is a fully gig-ified economy.
Alex Hanna: Yeah, absolutely. And we’ve seen this gigification happening in higher ed, in nursing for quite some time now. And so if we understand AI as an intensification of those processes rather than something that’s kind of brand new, I mean, I think that helps our analysis quite a bit.
Kelly Hayes: So there is a ubiquitous message that resisting AI is futile, and workers simply need to upskill and learn how to use it. But the Luddite Lab argues that some of these systems actively de-skill people, turning skilled work into supervising and repairing bad machine outputs, as you’ve been describing, while undermining opportunities to learn through doing work itself. Can you talk about what this rhetoric of reskilling gets wrong about what’s happening to skill, expertise, and professional knowledge?
Alex Hanna: Yeah. I mean, that’s a thing that I think is a bit worrisome. I mean, it’s not just worrisome in the workplace, but it’s also worrisome as we’re encountering students as educators. I haven’t been in a formalized classroom in a couple of years, but I’ve been an educator since I was an undergraduate about 20 years ago. Now I’m aging myself a little bit. And so the preliminary studies that we’ve seen on things like cognition, on retention when using these technologies is pretty upsetting. We’ve seen at least one study that suggests that cognitive ability, when measured by different cognitive imaging tools, seems to go down. There’s a study of a lot of Chinese students where I think they showed that, I think it was an experiment, students that used LLMs [large language models] to complete homework had higher homework grades, but had much lower test scores, which seems to signify that retention is going down.
And there is a study that was done by Anthropic themselves that showed that using AI-assisted coding tools, when people were learning how to use a new software library, they basically had no knowledge of how it worked or how you could execute with it. And so that could lead to things like cybersecurity attacks or not being able to check the code, not being able to understand what it’s doing in a way that is careful and thorough. And so this kind of thing is also okay with a lot of the tech executives. I haven’t read the whole of this interview, but it came out today. So Ezra Klein did an interview with NVIDIA CEO Jensen Huang about AI making kids worse at math and reading. And he basically said, “Well, I don’t think it matters.” And who cares if it’s going to be the same like people using calculators right now?
But if we’re thinking about calculators doing something of that nature versus doing any critical thinking, I mean, that’s very, very concerning as an educator. That’s concerning to me as a writer, as a reader, I’m very worried about that. And so, even if this is not the intention of tech leaders or corporate execs or corporate managers, but it does look like this is an intentional strategy to de-skill people and make them wholly reliant on the technology.
And so I’m very concerned that that is the future that is, that we are being sold. And on some degree, it is an intentional strategy to remove agency and to remove a certain kind of specialization of work, that we have gained these skills through hard-earned education and trial and error and through long years of education and experience. Now, can they do that in many fields where the automation isn’t quite there yet? Well, maybe it doesn’t work very well, but corporate managers may still argue that these are sufficient replacements and then that does lead to a further de-skilling.
Kelly Hayes: Absolutely, and this is something our listeners may be tired of hearing me rant about, but I really want it burned into people’s brains that Sam Altman thinks intelligence should be a utility that we have to pay tech companies to access. And you know, I think Marc Andreessen saying that he doesn’t believe in introspection is maybe also a bit relevant here. As a populace, we are being strategically simplified. And I don’t think you can really fight that unless you acknowledge that it’s happening.
One of the objectives of that simplification, in addition to, I would argue, pacifying the populace more easily, is making us more interchangeable as workers. So we have to consider what that means in the context of the gigification of our work lives. I know some people who aren’t tech bros or billionaires who say things like, Well, this is just what capitalism has always done, or what’s the difference between this and a calculator, and I think we really need to reckon with the extremes we’re dealing with, in terms of how the harms of capitalism can be multiplied and amplified with this technology.
Alex Hanna: Yeah. No, completely agreed. And I think this is helpful to also think through because when we hear somebody like Altman saying, “Well, we need to make intelligence a utility” — or the latest incarnation I’ve heard of this is Mark Zuckerberg saying that we need to democratize superintelligence. And what he means by that is now we have this news agent that pops up when you use WhatsApp or Instagram. It doesn’t really mean anything. It means that there’s a product that’s now on your other product, now on your other Meta products. But this trajectory of gigification is something that we’ve seen for … I mean, we’ve seen it for years. I mean, it has been popularized by Uber and Lyft, but the kind of history of the gig is temporary work, it’s contract work. Now that work is mediated by an app, it’s mediated by the platform. And so one of the things that we can see is a way that gigification and casualization is intensified.
I can speak specifically to the way that we see that in, let’s say, education. We already have this trajectory in education in which universities have really leaned into the casualization of teaching. I mean, the last time I was a professor, over half the classes in my department were taught by contingent faculty, by adjuncts. And that’s just increasing. I mean, tenure-track jobs are becoming more and more scarce, and more of this labor is being done by people who are being paid piecemeal. And so we can see a world in which, if the individual instructors are already seen as relatively interchangeable, that are seen as just the sum of their outputs, then an LLM-based university is just a dream. And some universities have really leaned into that.
A good example of that is Arizona State, who have this partnership with, I think it’s called GSV, they’re a huge ed tech firm. And so they have this partnership where so much of what’s happening at that university is that they are very much leaning into being this AI-first university. We see this intensely in higher education, in a place that already suffers from what the economist [William] Baumol called cost disease. The idea that in other areas, especially in manufacturing, the marginal cost of the widget goes down due to increases in technology or automation.
But the marginal cost of your violin player does not go down because people go to the symphony and that is something that requires a huge amount of training, but universities feel like they might be able to solve this cost disease with the usage of AI tools. I owe this analysis a lot to Dan Green, who has a great essay that I shout out every time I can, which is What Does AI Do? Specifically, it’s in the AAUP’s [American Association of University Professors’] publication Academe, but specifically why universities are leaning so much into AI. And so yeah, I mean, this kind of casualization, gigification, platformization, I mean, this is the way that LLM or synthetic texts and image generation slots into these fits very well and it points to a real secular trend.
Kelly Hayes: You know, speaking of academia, I was actually chatting with a professor recently who was talking about the complexities around students using AI, and how they had a lot of sympathy for students who felt drawn to it. Because they work at a school with a lot of working-class students who just haven’t had the same access to skill-building, in terms of sharpening their writing skills, that some more resourced students have, and they feel like the chatbots allow them to level up, and even out the playing field. This professor was really sympathetic to that position, and was actually one of the people who asked me: How is this technology so different from the way capitalism has always worked? And I do want to acknowledge that the inequities being named are real. Those inequities have been part of my own experience. We obviously need a lot of transformative change, in terms of the educational opportunities people do and don’t have as young people, and in general. But what’s really worrisome to me is that this technology isn’t actually leveling anyone up. It’s actually grinding everyone down, and sort of slapping a veneer on that process.
Real writing is valued less, or looked at as suspect, in a world awash in slop. And, under capitalism, producing slop doesn’t ultimately make a person more valuable — it just makes them more interchangeable with other people, who are also being devalued. And as we’ve discussed, it also cuts people off, in terms of their actual potential to gain ground with their skills. There are things I didn’t learn when I was in junior high and high school, because I didn’t go to the best public schools, and there’s so much I learned in my first year of college, from my favorite philosophy professor, about how to think critically, and how to scrutinize the values that were being normalized around me. I don’t blame students for using these products, and I don’t blame workers for using them. I think a lot of people are just trying to keep up, or don’t see the harm, or that they’re trying to level the playing field. So when universities embrace these technologies, it really feels like an institutionalized devaluation, just across the board, of academic workers, of students and their potential to learn, and of the purpose a place of learning is supposed to serve in the first place. And it really feels like an implosion of the university, conceptually, that’s been underway for some time, but is being supercharged by AI, and I think we should also think about who benefits the most from that destruction. Who views colleges and universities as the enemy? It’s not the left.
So, I am really with that professor, in terms of not being judgmental about why individuals would gravitate toward these tools, or punitive about those things, but I am deeply concerned about what people are being robbed of, and about how we’re all being devalued, collectively, in ways that I think are going to have really ugly consequences in the long and short term, unless we fight.
So, I really appreciate the Luddite Lab as a resource that helps workers think about how to fight back. I think, given the way this technology is being thrust upon us, it’s hard for some people to envision that fight. And some folks may imagine that it’s an all-or-nothing struggle, where we’re either banning the technology or subsumed by it. But the wins collected by the Luddite Lab are much more concrete protections against AI-driven layoffs and wage cuts, and involve things like human editorial control, disclosure requirements, and the right to scrutinize deployment. Can you give us some examples of what these wins have looked like across different sectors?
Alex Hanna: Yeah, absolutely. So a few different wins that we’ve seen. Let me start with journalism. We’ve seen several wins in journalism. One of the organizations that we profile is the Ziff Davis Creators Guild. So this is writers and editors that are affiliated with the company Ziff Davis. So that includes publications like Mashable, CNET, ZDNET, these are online only, online first news sites. And so they have contract language that allows them to have control of any AI-generated content, any synthetically generated content that would go on their site and to have them have editorial control as well as establishing an AI committee in which they can discuss these changes. So they were able to win that and came within the precipice of going on strike to win that contract protection.
Another example within journalism is a recent arbitration win that we saw with the PEN Guild. The PEN Guild is the union that represents about 225 writers and non-supervisory editors at Politico and E&E News. In the latest contract that PEN Guild had bargained over, it required that any new technology be brought to discussion with the union within 60 days. Politico ran afoul of those rules, did not advise the union and implemented this AI tool, which they called Capitol AI for Politico Pro members, which is a paid product that Politico has. And they also had a number of different AI summaries that they used during presidential debates. And this was not brought to the union’s attention. The union filed a grievance. Politico disagreed. They took it to arbitration and PEN Guild won. And they basically said this runs afoul of the contract.
These things need to be bargained over with the union. You need to have some kind of oversight over this if this new technology comes into play. So that’s in journalism. I talked a little bit about the New York Times Tech Guild and about the kind of layoff protections, the just cause protections. You brought those up, Kelly, and the establishment of the AI subcommittee in their contract that they eventually won with a short strike. And then one we haven’t brought up in this conversation is around National Nurses United. National Nurses United is a union that’s a super union of many different state and local associations. I live in California, so here it’s California Nurses Associations. In New York, it’s the New York State Nurses Association.
Their wins have had some contract provisions, but what really impressed us in speaking with National Nurses United is the depth of political education that they engage in. You see an upskilling of people within that union that educates their members about what AI is and the way that it shows up in their job. Because, Kelly, one of the things that we find is that when you talk about AI, you have to set table stakes. We have to have that conversation we had earlier, which is: What do we even mean when we say AI? And that’s something that needs to be discussed from the jump, because how do you even spot what’s happening in the workplace? So in nursing, that looks like automated transcription. It looks like automated charting. It looks like automated scheduling. It looks like patient monitoring. It looks like all these little pieces that are often built into health care.
It looks also like platformization that nursing is going through as well, the kind of ways in which there’s these platforms in which nurses operate more like Uber drivers. And so what National Nurses United does really well is that they established a political education program from the ground up. They have a really robust political education department that presents materials and also gives those materials to local leaders to train up their membership. And they also use the machinery of professionalization committees that nurses have to go through as a condition of maintaining their licenses, and they use that as a mechanism to also deliver political education. National Nurses United is really unique in the sense that they not only [don’t] stop at talking about these issues of productivity culture, of layoffs, of the use of AI as a cudgel against workers, but they’re connecting the data infrastructure and the political economy of these companies to so many other areas.
Right now, National Nurses United has been doing actions against Palantir because they know that the data harvesting that’s being done, some of those systems have to be outsourced to be built by someone else. Sometimes it’s Palantir. In England, in the U.K., the NHS [National Health Service], their data warehousing system was run by Palantir. As you and I know, Palantir is wholly implicated in ICE [Immigration and Customs Enforcement] raids, in border terror, in terrorism brought upon our undocumented neighbors and friends and family, in wars throughout the world. They’re just implicated in so much. And so their political education program doesn’t just stop at discussion of AI in the workplace, but the ways in which the organizations are building this kind of surveillance regime that is implicated with state violence.
Kelly Hayes: So what lessons do you think we can take away from these campaigns about what workers should be doing and demanding?
Alex Hanna: I mean, these are very inspiring campaigns and what I take away from them is understanding that anybody can fight AI. You don’t have to understand how something works to understand that it’s hurting you. You don’t have to understand how a gun works to know that you can get shot by one. You don’t have to understand how a synthetic text extrusion machine could say that you are not fit to receive health benefits and know that you are not going to get health benefits that month or that year. And so one of the things that we can take away is that anybody can join this fight.
The second is that the most powerful way that we really are able to fight these is if we’re in a collective, if we are in a group of people. Of course, I’m channeling Mariame Kaba, thinking about anything worthwhile is done with other people. And I think that’s something that is very helpful. I think one of the things that we think about in any of these tech justice struggles is understanding who around you is already doing this work. Where can you plug in? Where does it make sense to? Is there a way that you can lend your own expertise or skill set?
And then I think with worker struggles specifically, one of the things I think is really heartening is that in the U.S., in an era of rising fascism where the government — the federal government, and more in many places, the state government — is absolving any responsibilities or ignoring any checks that can be put on these tools. Workers’ organizations are one of the strongest types of counter-governance structures that we have. So workplace democracy means real life democracy too. I mean, not that workers, workplaces are not real life, but workplace democracy reflects state governance, and in a vacuum, state governance as well. And so in that case, that means governance over technology, governance over what technology is doing in our lives and in our workplaces.
Kelly Hayes: The Luddite Lab also stresses that these fights begin long before bargaining and continue after a contract is signed. You’ve been talking about examples of strong political education, inoculating workers against the idea that adoption is inevitable. What does good political education around AI look like and what resources should people be tapping into?
Alex Hanna: Yeah. Good political education around technology and AI looks like one, doing a little bit of denaturalization of what we mean when we say AI. What does it mean to talk about these technologies? In that case, we often rely on sets of questions, like in a particular type of tool, what goes in and what comes out? Does it make sense to connect the inputs and the outputs? Who benefits when the system goes well? Who benefits when it actually goes not as well? What are the companies that are building these tools? What other contracts do the companies have? Where does the data go? Is there any way to protect the data? Is there a way to remove the data? Is there any kind of meaningful opt-out? What does consent look like in these cases? And then what is the supply chain in the development of this? I mean, what are the other data practices or what are the other labor practices that these companies are implicated in?
And so it is rarely about the technology. You don’t need to know a lick about machine learning or neural networks to understand the connections between inputs or outputs, whether that even makes sense on the face of it, but you can understand the kind of network of organizations and the ways that they are implicated in the workplace. So the most successful political education programs have been the ones that do that. And so again, I want to continue to point to National Nurses United that does a great job of talking about that, developing those tools, connecting it with other kinds of pieces of political economy that talks about the data infrastructure and the ways that these companies are implicated in border terror and war making.
Kelly Hayes: The word “democratization” gets thrown around a lot when people talk about AI and its outputs, which is actually really ironic, given the fascist environment we’re living in, and the ways this technology is being weaponized toward fascist ends. But if we took worker control seriously, what would a genuinely democratic politics of technology look like — not simply thinking about workers getting to say no to a bad system, but workers and communities having the power to decide what technologies are worth building, and what problems are worth solving?
Alex Hanna: I think it would look dramatically different. I mean, first, there’s some good discussion to be had about thinking about work more generally and to think about what kind of work that we want to do in life. And if we think work can be worthwhile, then what would it mean to have technology that actually facilitates people’s jobs? It can look like a number of different things. I don’t purport to have any kind of solution ready-made as I shouldn’t because I think so much of this solution is much more local and is much more bound in what local conditions are. So for instance, what could be other conditions of worker organization that are more cooperative in nature? If they’re more cooperative in nature, what might be types of technologies that facilitate democratic decision-making in the workplace? What are ways that people can actionably change those pieces of it?
There’s many different experiments in democratic decision-making in technology. What would it look like for, let’s say, people who are non-technical to contribute to technology projects? What would it look like to encourage people who don’t have, let’s say, coding skills to contribute in such a way in which their needs are taken into consideration, rather than that kind of knowledge being generated or produced by some benevolent tech dictator?
And I think we need a lot of experiments in this. There’s a person, Rafael Grohmann, he’s written a bit about platform cooperativism and the idea that there’s joint platforms where individuals are in tech worker collectives. And so that might be an alternative model for control over one’s workplace, which I think then extends to control over the technology. And so that’s one thing that comes to mind. But I really think that one of the things that I have an armchair theory about, which I think is correct because it hasn’t failed me yet, is that technology really reflects the organizations in which it is embedded.
This is not a new idea perhaps, and when I say it, it’s rather simple-sounding. But effectively, one argument that’s a mainstay of science and technology studies is that a weapon like the atomic bomb could not exist if it was not for a huge organization that had to manage all the accesses or do all the building of such a thing. And it requires a bit of an authoritarian structure of organization to maintain. And this is a point that Langdon Winner makes in some of his texts. The same could possibly be said of technology where if you have a collective, then it might look like your technological development is much more communal or collective. If you have something that’s wholly top-down, then it is fine to buy ChatGPT Pro as a service and to give it to all your employees, especially if ChatGPT allows an administrator to see everybody’s prompts and outputs. I mean, that’s a full panopticon of boss access. So I do think there’s that kind of reflection, that technological world we want to see also depends on what kind of organizations that we make and build as alternatives to capitalism and fascism.
Kelly Hayes: I appreciate these insights so much. As we wrap things up here, is there anything else you would like to share with or ask of the audience today?
Alex Hanna: My only ask or my only request of people is to know that you, no matter where you are with regards to how much you know about technology, have the ability to direct where technology should go. And it can be as easy as getting with other people in your community and understanding and trying to get a sense of what your needs are, and then articulating that in a way that you could possibly build something different and newer and better and more communal. So that’s probably my last optimistic thought.
Kelly Hayes: Well, Alex, I want to thank you so much for joining me today. And where are all the places people can find your work?
Alex Hanna: So you can learn more about the Luddite Lab at labor.dair-institute.org. And then you can follow me on Instagram at AI_Killjoy as well as on TikTok. And then I’m on Bluesky @alexhanna.bsky.social. And then you can just find me around. I’m fluttering around the interwebs.
Kelly Hayes: Wonderful. Well, I really appreciate you making time to join me, while you’re still recovering from not feeling well, and I really hope we can do this again sometime.
Alex Hanna: Absolutely. Thank you so much for having me, Kelly.
Kelly Hayes: I also want to thank our listeners for joining us today, and remember, our best defense against cynicism is to do good, and to remember that the good we do matters. Until next time, I’ll see you in the streets.
Show Notes
- Don’t forget to che
Post truncated. Read the full article at https://organizingmythoughts.org/ai-is-class-warfare-workers-are-fighting-back/.
From Organizing My Thoughts via This RSS Feed.