I’ve spent so many years among the Luddites—among oral histories, archived letters, and old newspaper articles about them—that there are scenes from their history that are so seared into my brain, it’s almost as if I was there myself. I can conjure the battle at Rawfolds Mill, where more than a hundred clothworkers made their ambitious and ill-fated assault on a hated factory owner who used automation not to ease their burden but to undercut their wages and employ child labor. The clothworkers, under the banner of Ned Ludd, were crushed; they left trails of blood in the mud as they fled into the forest.
It’s the quieter moments that are more indelible, however, and more likely to come to mind as I’m reading news of AI startups and artist strikes. Take the young Luddite leader George Mellor, remarking to his cousin on an evening walk that he’d seen how bosses used automating machinery and found “the tendency’s all one way”—to concentrate more and more wealth and power at the expense of workers. Mellor was, and continues to be, right about that, and he made this observation in 1811, before industrial capitalism was fully forged.
Or take the debate, between Mellor and an apprentice saddlemaker, John Booth, about how best to address the rise of the industrial entrepreneur, mechanization, and the factory-owning class. In an account first recorded by the historian Frank Peel, one that is surely embellished, Booth argues that the Luddites are right to resist the factory owners, but that they should embrace the technology—they should rebel for reform, not for refusal. “I quite agree with you . . . respecting the harm you suffer from machinery,” Booth said, according to Peel. “But it might be man’s chief blessing instead of his curse if society were differently constituted. . . . To say that a machine that can do this for you is in itself an evil is manifestly absurd. Under proper conditions it would be to you an almost unmixed blessing.”
“If, if, if!” Mellor interrupted. “What’s the use of such sermons as thine to starving men? . . . If men would only do as thou says, it would be better, we all know. But they won’t. It’s all for themselves with the masters.”
It is hard, even futile, Mellor argues, to imagine a world where an advanced technology is put to the common good when it is erasing livelihoods right now. If dismantling the machinery ends an injustice and tips the scales toward equality, that should be the project. This dispute, too, remains as relevant and pointed as ever, two centuries later. It lies, I believe, at the heart of the matter of actualizing the laudable prescriptions put forward in Morozov’s essay: reimagining how we might develop and institute technologies like AI more democratically, more holistically, more attuned to amplifying humanistic and scientific pursuits—and decoupled from its current death drive to profit management at any cost.
I find Morozov’s vision of an eolithic mode of technological development—one in which we are free to experiment with and develop technologies not to advance the aims of a military or administrative state, not to realize corporate efficiencies, perhaps not even toward any design at all—to be a beautiful, even moving, one. It says much about how severely Silicon Valley capitalism has narrowed our imagination that so relatively simple an idea can feel so utopian. I’m also in staunch agreement that we are in dire need of such reimaginings, and a concerted effort to make room for them.
The question remains how to get there from here, which brings us back to Mellor and Booth’s argument. The path will include a radical political project, as Morozov notes, and yes, meaningfully democratizing AI will require true small-“d” democracy. But a project of what character? Of radical resistance, or of political reform? Of revolution or abolition? The stylized opposition between “realists” and “refuseniks” may prove far less rigid in practice.
Again the Luddites might offer some wisdom. Generative AI presents a host of threats to working people. It promises to increase surveillance, exacerbate discrimination, and erode wages. It threatens to concentrate power among the Silicon Valley corporations who own and operate the large language models, and their clients among what Morozov aptly termed the Efficiency Lobby. These companies are accumulating investors and market cap, while hurtling toward IPOs that stand to richly reward stakeholders, regardless of whether the enterprise AI software delivers as promised or not. The tendency’s all one way indeed.
And once again, the workers most immediately—and perhaps most existentially—threatened by generative AI companies are skilled craft workers. Visual artists, writers, musicians, voice actors, and illustrators; journalists, copywriters, graphic designers, and programmers. Many of these workers have joined a campaign of refusal, of modern Luddism: class action lawsuits to try to stop the AI companies from profiting off of the wholesale, nonconsensual appropriation of creatives’ labor; efforts in organized labor to stop studios and corporations from using AI to generate scripts or animated productions; consumer campaigns to declare goods and services as AI-free. A still-underappreciated truism of our technological moment is that there is great solidarity to be found in refusing a technology—AI, mostly—that is used to exploit or replace a worker.
The striking Writers Guild of America (WGA) screenwriters, whom I spoke with on picket lines and at rallies, were surprised to see their cause become so celebrated in 2023, when they drew a red line at allowing studios to use AI to generate scripts. They ultimately won the right to use AI how they saw fit in their own creative process, ensuring that, for three years while the contract holds, at least, if they use AI at all, it will benefit them, and not their bosses. Here we might see the seeds of a potentially radical project to move control of how a technology is used into the worker’s own labor process—born of a refusal to allow management to claim that right for itself.
The WGA is a uniquely powerful union in the creative industries, of course; many other jobs, including illustration and copywriting, are often freelance and more precarious. Any movement will have to work to encompass these workers too, as well as the numerous data cleaners, quality assurance testers, and content moderators on whose labor—carried out in stressful conditions for abominable pay—the whole system depends. And I think we must recognize that democratizing technology means offering access to a kill switch—and that generative AI, in its current formation, may well be deemed too wasteful, too undermining of the creative trades, too polluting of the information ecosystem, and too toxic to stand. Booth ultimately joined Mellor’s Luddites, after all.
Like him, however, I see that many of AI’s ills stem from those who control and stand to profit from it. A democratic movement might equally well cut off the plagiarism and slop production and redirect this technology toward predicting new proteins and writing custom language apps. The key to achieving any alternative routes, eolithic or otherwise, will lie in the scaffolding—in locating productive ways to harness the energies and solidarities of refusal into a broader project of reclamation, of reimagination, of renewal.
