A little over two centuries ago, the Luddite movement of English workers took a valiant stand against technological disruption. They broke new machines that put them out of work, they sang songs and sent threatening letters, they demanded security and some measure of control over their own lives. And they lost. 

Today the Luddites are a touchstone of left political discourse and part of the ongoing public conversation about AI. I get it– ChatGPT has invaded my classroom and is wreaking havoc in education at all levels right now. I don’t even like DLSS, the Nvidia technology that increases frame rates in video games by predicting likely frames rather than fully calculating each one– this kind of frame generation looks bad to my eyeballs and feels like a shortcut in place of proper optimization. (This is a cranky opinion that will age poorly. I hold it anyway.)

But whatever you think about AI, the example of the Luddites doesn’t offer a response that’s either viable or desirable. Elsewhere in the history of early 19th-century artisan movements, we can find a better alternative.

Artisans were a kind of worker unfamiliar in the modern economy. They were not replaceable proletarians at the bottom of the economic ladder, like 19th-century factory workers. Nor did they have the institutional arrangements or formal education of modern-day professionals. Rather, they had specific skills learned typically through apprenticeship and guarded by guilds that gave them power in the early modern economy. Blacksmiths, shipwrights, tailors, and other artisans could do things that few others could do. 

There is a lot to defend in this world! There is also plenty to condemn. Artisans brutally exploited their apprentices. In British shipyards, apprentices were regularly beaten; the 18th-century apprentice William Chandler learned remarkably little from several of his masters, and instead had to educate himself in between doing their dishes and ducking their violent family disputes (which included, at one point, tossing an infant). In some industries, artisans’ wives and children provided critical unpaid labor, and artisan culture could be deeply patriarchal. 

With the rise of new machines in the late 18th and early 19th century, artisan life began to change. This happened in waves: some textile work was mechanized as early as the late 18th century, while it would take decades into the 19th century for wheelwrights (the guys who make wheels) to face significant disruption. But industry by industry, wealthy men (sometimes well-off masters within the trades) took artisans’ knowledge and built it into machines that made work simpler, faster, and easier to supervise and control. The basis of artisan power had been artisan knowledge and economic value– with new machines, fewer and less skilled men could do artisan work. 

Work itself had changed: gone were the days of small artisan households handcrafting their way through production. Now work was becoming a series of isolated tasks. An artisan’s day was under his control (if he wasn’t an apprentice or, in some cases, journeyman) and intellectually challenging. Potters who had spun and thrown and decorated their wares were replaced with assembly line-like setups of men who did one part of the complete process again, and again, and again, and again. They generally made far less money and worked far longer.

It should be little wonder that workers resisted violently. An increasingly common boiled down version of academic scholarship on the Luddites holds that they were not anti-technology, but rather broke machines as a negotiating tactic. We might say they wanted technology on their terms. This is partly true, but it’s also hard to assess. Across industries, new technology was implemented specifically to disempower laborers. When unknown saboteurs destroyed steam engines in Britain’s Royal Dockyards, they were resisting both new machinery and the new agenda it came with. They did not leave notes to tell us what policies they might prefer in a vacuum. Early 19th-century machine breakers at least appear as frequently hostile to technology itself, but that hostility is inseparable from hostility to its effects. This shouldn’t be surprising: they experienced new technology as a series of intrusions and impositions, not as a high-level economic phenomenon. 

We can class the Luddites with a broader artisan political tendency in this era that, regrettably but for good reason, frequently looked backward rather than forward. One example worth considering is British naval shipwrights. Over the course of the 18th century they used strikes to vigorously defend their privileges– limited control over their workday, the ability to take waste timber from dockyards, a relative lack of supervision– from administrative encroachment. As Peter Linebaugh tells in his book The London Hanged, this artisan world was finally destroyed when Samuel Bentham remade the actual work process itself, including with machinery, in the 1790s-1800s. 

===

This was not the end of the proud artisan politics of the 18th century though. While there is tragedy, this story does not have a tragic ending– not if we narrate it right. In 1797, a young shipwright named John Gast was fired from His Majesty’s Royal Dock Yard at Portsmouth– perhaps for helping successfully negotiate a labor dispute and getting a supervisor fired, or perhaps because layoffs generally prioritized younger men first (I told you artisan culture had issues). Gast went to work in the more precarious private shipbuilding industry on the Thames River, where he became a successful organizer, ran a pub, and preached, among many other things.

John Gast did not, and could not, resist technological developments that had already happened. In the 1820s he published a number of timber reform plans quite similar to Samuel Bentham’s. He also collaborated with middle-class reformers to develop new adult educational institutions called Mechanics’ Institutes. Gast insisted that working people could and wanted to learn the latest in science and technology– and politics and economics– to exert power in the future. New technology was often developed as a a weapon meant to destroy working people, but they could fight over who was holding that weapon. 

As Linebaugh puts it, “it was the genius of John Gast” that he did not question the form of new economic relations (Linebaugh is focused on wages, or “the money form” here; I think it’s broader than that, but that’s an issue for my book) but instead “adapted forms of struggle appropriate” to the new present. Gast built mutual aid societies into an early shipbuilding union, which he defended against legal assaults, including with brilliant and arguably deceptive testimony before Parliament. He became a member of the “radical” movement for democracy, working roughly in between popular radicals in the streets and the moderates who were represented in Parliament. I can tell you about him because all this prolific writing and organizing left a paper trail for historians like Iorwerth Prothero and Peter Linebaugh (and me) to follow. 

What he did not do was blow up machines or try to restore what Linebaugh calls the “art and mystery” of prior shipbuilding (I sense some critique from him there, but I’ll save the historiography for the book). Economic historians like Joel Mokyr have, infuriatingly, accused workers of resisting technology because they want to protect the “rents” their knowledge generates. Artisan knowledge was labor value, not rent, and they were proud of it. But if artisans had somehow forced everyone to forget new machines were possible, or broken every single one before it was made, they would have been rent-collectors. That was not only possible, it was undesirable. 

Many critics of AI– myself included until recently– rest comfortable in the notion that AI is useless or a scam. Even if this were true right now– and I think the transformative role of Claude Code in software development over the last few months makes this undeniably wrong– there is no guarantee it would be true forever. If it’s possible to automate any labor, capitalists will try– must try, thanks to competition– to do so. If you base your goals for the future around what is true about the world, you are asking to be defeated. We have no control over what the physical laws of the universe allow.

Maybe AI will plateau. I refuse to bet our futures on that. There is no unlearning the fact that you can throw language and linear algebra in a bag and shake it around and get useful results– systems that can program computers and generate images and do plenty of useful work as well as plenty of useless and evil work. But we can dictate the conditions within which these new tools are used. 

We have to organize around how things should be, not how they are. Here’s a grab bag of policies that could include. First, shorter working hours. The easiest way to ensure productivity gains go to all of us, and employment does not crater, is to give us more time to do what we want. Second, educational siloing. We can restrict the use of AI in an educational setting to places where we determine, collectively, that it’s useful: kids need to learn how to read and write without ed-tech consultants tossing new technological obstalces in the way. “AI in education” should mean giving high schoolers linear algebra, not demanding social studies teachers have them generate essays instead of writing them. 

Third, de-enclosure (which might manifest as nationalization). AI only works because of our humanity’s crowning achievements enshrined in language itself. We did that together. We created art and music and literature and programming. Any technology built on the back of that should be shared. This could look like government ownership, open source, democratic controls– what it clearly does not look like is private for-profit development. 

Fourth, regulation. By enforcing existing law (Grok generated child pornography for weeks, which is illegal, and “oopsie!” is not a legal counterargument; Elon Musk should be in prison) and passing new legislation, we can make many of the harmful things one could do with AI illegal. A broader regulatory framework could then include oversight mechanisms for maximizing benefits as well as minimizing harms. It’s not possible to ban AI, but it is possible to punish generating misleading videos, for example. 

These are just policies though: we don’t just need a program, we need strategies. I don’t know what they are yet. What I do know is that they will be different from what came before. Gast grew up in a world of mutual aid societies and informal organizing in which most modern labor organizing would be illegal. He helped change the law and developed new ways to organize workers in their workplaces and in politics. Some of his new ways are now our old ones. Unionizing, running for office, calling legislators— these are important, but won’t be enough for even the limited program I just sketched out. Much of the wishlist I just wrote sounds impossible to even me— but that’s just because we haven’t figured out how to make it possible yet.

In short: accepting that AI exists does not mean accepting where or how it will exist. All too often we here “it’s inevitable, it’s coming!” as an argument for passive acceptance of AI in our elementary schools, our workplaces, our lives. But acknowledging that economic conditions are changing should mean taking a proactive role in determining how they will. We can do this, together. That requires us to reject the easy solutions offered by billionaires– just accept it and make us money– but also the easy apathy that dismissing AI’s capabilities encourages.

John Gast did not get everything he wanted. But he got more than most people could have dreamed. The radicals succeeded– in 1832, parliamentary reform achieved through the maneuvering of moderate radicals and fear of popular radicals made Britain more democratic than it had ever been. In 1833, Parliament finally abolished slavery– because the new democratic majority was less hostile to doing so than the former oligarchy had been. Gast’s unions grew and evolved into one of the most powerful labor movements in the world.

In 1750, government shipwrights were among the few workers in the world to have public healthcare and retirement pay. Within two centuries, thanks to the workers’ movement Gast helped build, British working people achieved national healthcare. (And today, wooden shipbuilding is a rare but valuable trade primarily done by people who love it, rather than miserable apprentices dodging falling timber and ducking the blows of their masters.)

This was not because of Luddite agitation or backward-looking artisan appeals. It was because workers took up the hard challenge of reimagining their economy and politics after both had been forever changed, unjustly and unpredictably, by people seeking to destroy their way of life. The best of the artisan vision of politics is alive today. We can honor it by doing what John Gast did: not trying to wind the clock back 10 or 20 years to an imperfect past, but charting a new path to a future on our terms. 

You can read more about John Gast in Iorwerth Prothero’s partial biography of him, and in this 1802 pamphlet Gast wrote defending a strike by shipwrights.