It seems that nothing could be more popular than AI. But also less popular. The enthusiasm people exhibit for letting machines do their thinking for them, from taking notes to writing memos to cutting out the joy of the creative process in making art, is equalled only by their hatred of “customer service” bots that prevent customer service and giant AI data centres that suck all the power and water from the Earth and leave it in desiccated darkness. As Heatmap notes, “Americans have swung en masse against local data center development. As recently as August, Americans were split on whether they would support a new data center built where they live; now, 70% would be against it.” A New York Times columnist just urged Democrats to jump on the anti-data-centre bandwagon because “According to polling by Heatmap News, more than half of all Americans support a national ban on data centers. The public seems to agree that data centers are giant, ugly, noisy, smelly altars to industrial-scale hostile architecture. In our virulently partisan country, this constitutes a rare show of consensus.” It’s not clear exactly why it being local really matters since these things are not especially loud and do not belch thick smoke or suck careless passers-by into turbines. There are legitimate concerns including, of course, that their power demands are pushing up everyone’s electricity bills, although the increase is partly due to local politicians having spent years trying to shut down generating capacity. But as Heatmap also points out, opposition to data centres is part of rising opposition to AI overall. Which is an important further wrinkle. if you’re using Claude Siri Copilot Whatever six hours a day, where do you think the computing power is going to come from?

It’s not entirely clear. As Heatmap also points out, opposition to data centres is in part opposition to AI overall which a lot of Americans now think is itself, on balance, a harmful development:

“45% of Americans are ‘pessimistic’ about AI’s effect on their lives – and 55% are downbeat about its effects on ‘society as a whole.’ Young people are particularly downcast. In virtually all of our polls, American adults younger than 34 stood out for being opposed to data centers and AI. Meanwhile, the only group that’s outright optimistic about AI’s effect on their lives? Men older than 65.”

What the latter are thinking we can’t imagine, unless it’s that they’ll get a robot nurse then die before the headless robot cheetahs with laser beams start rounding up humanity. Arguably a narrow view. But it’s also possible that people who actually learned to think before the stuff showed up are better-placed to use it as an aid to thinking rather than a substitute.

It also seems clear, a phrase here meaning nobody but us appears to think it, that it is precisely the things people like least about AI that are most likely to stop it from doing what they fear most. Reuters, for instance, warns that:

“Data centres are expected to consume twice as much power and water by 2030 as they expand to meet the surge in demand from artificial intelligence, U.N. researchers said on Wednesday.”

Of course they then do the usual big-government-saves-us thing:

“Unless governments heed the rising environmental ‌costs of AI, the rapid rollout could also strain scarce land resources and create mountains of electronic waste, the United Nations University Institute for Water, Environment and Health warned in a report.”

The UN what? And if they knew governments like we do, they’d be very afraid that the policy response would be to shunt water away from annoying dirty organic people to gleaming data centres. Especially given all the hype about how if we don’t keep up with AI we’ll be loser chumps in the global race not to be loser chumps. Which has a bit of plausibility because of the military applications.

In any case, it is clear that data centres are very power- and water-hungry, and increasingly unpopular in countries where popular opinion has any influence on governance. So not in China. But so unpopular in democracies that, as we noted recently, some people including Elon Musk are thinking of putting them in space where the neighbours won’t mind. (Or if they do they’ll have tentacles and highly acidic blood and vastly superior weapons so… uh… wait a minute…) And presumably it being cold in space you don’t need all that water to cool them.

Here on Earth you do. It’s not an illusion. They require immense resources to replicate what our brains do in a weird way. Of course people don’t always know what to do. As another New York Times writer recently kicked off her column:

“The backlash against energy-guzzling data centers has snowballed in recent months, prompting more than 100 proposed moratoriums at the local, county, state and national level. Proposals for data center moratoriums fall into two categories: Most call for a short-term pause in permitting or construction to give local governments time to establish guidelines. A few others call for permanent construction bans.”

A short-term pause to give governments time to establish guidelines. Now there’s a plan nobody could love. Guidelines for what? Making data centres as nice as strip malls? Maybe ask Alexa for a suggestion. Or Canada’s Prime Minister Mark Blarney, since the Toronto Star “First Up” email newsletter for June 5 assures us that:

“Mark Carney’s long-awaited AI strategy directs billions to building ‘safe, reliable, and sovereign AI.’”

As we’ve noted, Carney is a past master of the “glass bead game” of stringing together syllables in glittering abstractions. Safe is good. Reliable is good. Better than unreliable, anyway. But what is “sovereign AI”? And how will the typical Canadian government habit of squandering vast sums it has to borrow make AI “safe”? And could Mark Carney switch the input port for an email service if the fate of humanity depended on it? What does he know of “recursive self-improvement” in large language models and what could he do if he did know? Let alone the clowns in his cabinet.

His own press release, a gleaming cosmic abstraction, includes such prose as:

“While Canada has world-class talent and one of the fastest-growing digital sectors in the G7, we are among the slowest countries to adopt AI at scale. This gap risks undermining public trust, driving Canadian talent and startups abroad, and leaving critical parts of our AI ecosystem under foreign control. With the global AI market projected to reach U.S.$4.8 trillion by 2033, Canada has a limited but real opportunity to ensure AI works for all Canadians – to harness this technology to create jobs, protect Canadians, and strengthen our prosperity. ”

Works for all Canadians. Gosh. Swoon. And so they’ve leaped into committee:

“Today, the Prime Minister, Mark Carney, launched AI for All, Canada’s new national AI strategy. Over the next five years, this strategy will introduce new legislation, investments, and programs that ensure AI is adopted responsibly, in a way that truly serves all Canadians – building trust, expanding opportunities, and reinforcing control of our sovereignty. The AI for All Strategy targets an additional $200 billion of economic growth to create 250,000 new AI-related jobs over the next five years and to increase AI adoption from just over 12% to 60% by 2034.”

AI for All. Gosh. Swoon. But what does “targets an additional $200 billion of economic growth” even mean? What could it mean? Nobody knows. Not even him. Meanwhile you might want to keep thinking for yourself.

BTW Heatmap also opines, bafflingly, that:

“AI IPOs Could Create a Wave of New Funding for Climate Tech/ All that cash has to go somewhere. Why not philanthropic funding for decarbonization?”

Uh because it’s going to go into building data centres? All this power doesn’t consume itself, you know. Though if the tech bros who once embraced climate alarmism then backed off fast when they realized it was going to stop them from getting even richer and more powerful are feeling guilty, groups like Heatmap might be able to leverage some sweet donations.