Anti-AI feeling is real, rising and pointed squarely at the wrong target

If you spend any time at all on social media in the middle of 2026 you will run into the “AI is the antichrist” type of post. Data centres are going to drink America’s reservoirs dry and suck the electricity out of its cities. AI is enfeebling our brains, such that generation gamma will arrive with smaller skulls. That sort of thing.

Some of the feeling underneath is real enough. Gallup polling, reported in the Financial Times this week, found that nearly three in four Americans don’t trust businesses to use AI responsibly, that more than seven in ten don’t want a data centre built anywhere near them, and that eight in ten or more won’t trust AI to drive a car, offer medical advice or decide who gets hired. The New York Times has called fear of AI the most bipartisan issue since beer.

What I keep wondering, though, is how much of the hatred is about the technology and how much is being stirred. Spend long enough in these comment threads and you start to suspect you’ve met some of these accounts before, back when they were campaigning against 5G masts. They’re certainly the same crowd sharing the fossil-fuel-scented “1976 was hot and we survived” memes every time there’s a heatwave. [For the record: during the hottest fortnight of 1976, deaths in Britain ran roughly 20 per cent above normal, and past 30 per cent in Greater London at the peak. The surviving was not universal.] I’m not saying they’re all flat-earth, anti-vax, moon-landing-filmed-in-a-shed types; it’s unwise to stereotype. But most people don’t have original opinions. They repeat what they heard from someone louder, richer or more popular. [The internet insists Dostoevsky said that. He didn’t, which rather proves the point.] Either way, these claims shouldn’t go unchallenged. There’s an information war on, and facts, as usual, are the first casualty.

Andrew Neil claimed in the Daily Mail recently that US intelligence believes China is already manipulating American opinion against the construction of AI data centres. Perhaps. Or perhaps that’s American propaganda, designed to demonise China and hand big tech a hall pass to carry on funnelling unprecedented quantities of other people’s money into a technology that might cure cancer, or might build a surveillance state that never blinks, or conceivably both before lunch. That’s the trouble with propaganda these days: there’s so much of it about that asking cui bono doesn’t get you as far as it once did. Following the money is inadequate. Follow the power instead.

Either way, it’s a dangerous game everyone is playing. The firebombing of Sam Altman’s house in April was only the visible tip of a spate of attacks on people and infrastructure connected with AI, serious enough that the International Centre for Counter-Terrorism has published an analysis of it as an emerging form of anti-technology political violence. [The man charged with the Altman attack had spent time in anti-AI communities online, but investigators established no link to any organised group, and both Stop AI and Pause AI denied he was anything to do with them.]

Still, the shape of this is uncomfortably familiar. In the late sixties, opposition to the Vietnam war swelled the ranks of Students for a Democratic Society, a peaceful campus organisation that had existed since the start of the decade. Out of SDS split a faction called the Weathermen, who had decided that marching wasn’t working. They bombed government and corporate buildings, phoning warnings ahead to clear the people out, because they had concluded that the system behind the war was too violent to be reasoned with. As Bernardine Dohrn put it: “There’s no way to be committed to non-violence in the most violent society that history has ever created.”

You can see where the parallel is heading, and it’s worse this time, because in the sixties nobody at the Pentagon was issuing press releases agreeing that the war might end the world. Today the corporate sponsors of AI say exactly that about their own product, in interviews, cheerfully. If you took them at their word, the Weathermen’s logic would look almost restrained.

Which is precisely why the word should not be taken, and the logic should not be followed. It’s worth remembering how the Weathermen’s story actually went. The first people their bombs killed were three of their own, in a townhouse in Greenwich Village. The next casualty was public sympathy for the anti-war movement itself, which the American state gratefully accepted as a gift. Violence didn’t shorten the war. It handed the people running it a better story.

And it would be an even better gift now, because the anger is already aimed at the wrong thing. The Luddites understood this about their own machines, whatever history says about them: they weren’t at war with looms, they were at war with the factory owners using looms to break them, and the owners won partly by getting them written up forever as men who hated machinery. Broken looms made a much better story than broken wages. The same trick is running again, and it’s working beautifully. Nobody rattling a data centre fence is asking the question the owners actually dread.

So let me say the simple thing plainly, because it’s the only thing worth repeating. AI is not the enemy. AI is maths, and maths has never hated anyone. The enemy is the small number of people who own the maths, who bought the politicians now refusing to regulate it, and who would like nothing better than for your anger to spend itself on a data centre fence or somebody’s front porch. Every petrol bomb makes their case for them. Every burnt house buys them another year of doing exactly as they please. The way through is older and much duller: law, tax, unions, votes, and saying the same true thing over and over until it sticks. AI is not the enemy. The people aiming it at you are. I intend to keep saying it until it’s boring, which is roughly when propaganda starts to work.