Transcript: How Opposing Data Centers Can Save Democracy

Writer Astra Taylor argues the growing opposition to data centers is a huge opportunity for the Democratic Party.

This is a lightly edited transcript of the June 12 edition of Right Now With Perry Bacon. You can watch the video here or by following this show on YouTube or Substack.

Perry Bacon: We have a great guest today, Astra Taylor, one of the smartest people I know. She’s done documentary filmmaking, she’s written a ton of books, she’s an organizer with the Debt Collective, and she’s a person who’s just studied and is very thoughtful about a lot of different subjects. So Astra, thanks for joining me. Welcome.

Astra Taylor: Thanks for having me. Glad to be here.

Bacon: So we’re going to start with the topic of the year, millennium, decade. I want to talk about AI for a bit, because you wrote a piece I’m interested in. The title’s in The Guardian: “The fight against data centers isn’t just about tech, it’s about democracy.” But let me start with a basic premise here, which is: are data centers inherently bad, and is AI inherently bad? So talk about those things first.

Taylor: Oh, those are some big questions. Are data centers inherently bad? No. And data centers aren’t new. They’re new in the news, right? But data centers, 20 years ago, before we were talking about AI—data centers are where we store our data, and we were storing our data for old-fashioned social media usage or other streaming services.

So data centers have been around for a long time, and there was a big boom, a data center build-out during COVID actually, when internet usage exploded and there was a lot of access to low-interest capital that facilitated the build-out.

And one way of thinking about data centers is they’re the backbone of the internet, right? It’s where the cloud comes to earth. But they’re obviously much more prominent now, and they’re just being built at a different scale—hyperscale, to use the term.

Bacon: Let me come back to that, though. Data centers themselves have existed a long time. That’s what I wanted to get at.

Taylor: Yeah. So they’re not inherently evil.

Bacon: It’s new in the news, but it’s not new. We’ve had data centers. That’s what I was trying to draw out a little bit.

And I guess I do want to ask—a lot of people, I’ll say on the left, are very AI-skeptical. And I wonder—I think we can talk about the economics of it and the growth of it, but is AI inherently bad itself? It’s a very broad question, but I’m just curious what you think.

Taylor: I think AI in this economic model, in this political economic paradigm, is veering towards inherently bad. You cannot separate the technology from the economics, and this is a point I’ve been making since my first book, which is called The People’s Platform: Taking Back Power and Culture in the Digital Age, which came out in 2014. That was essentially a political economy of the internet, of the old-fashioned pre-AI internet.

And my argument there was that you cannot separate technology from the underlying business model of these firms. And that just seems to me like one of those basic eternal insights we should not lose sight of. And so in a sense, this is the same movie but just on steroids, right? The AI boom is happening in a period of much more intense wealth concentration.

So the “inherent” question—people like to say technology is neutral. I think that’s a bit wrong. Yes, you can use machine learning to assist a robust scientific infrastructure, or you can use machine learning to enhance a drone that is engaged in a genocide.

Bacon: That’s what I’m getting at. Could there be a world where AI is used nicely?

Taylor: Technology is flexible, yeah. But we would need very different societal conditions and be operating under a different government with a lot more constraints. And I just want to say on the “neutral” point: yeah, sure, you can use a knife to kill someone or to make a sandwich, but that doesn’t mean it’s neutral. It’s a tool that cuts things.

And I think the AI that is being designed right now is being designed for specific purposes. OpenAI—the definition they have of AGI, artificial general intelligence, that they’re looking towards is a tool that can do economically valuable labor. In other words, they’re trying to build a human worker replacement engine. And so neutrality—I don’t think this technology is neutral, but I also don’t think it is inherently good or bad. It’s embedded in societal conditions.

Bacon: So we’ve talked about data centers and now AI. Now we’re talking about AI data centers. How did this happen? You’re in North Carolina, but I think it’s nationwide. I feel like in the last 18 months, you’ve had AI data center protests, bans, really almost every part of the country—rural, suburban. Really not urban, because data centers are mostly in more spread-out areas, but how did this happen?

Taylor: Yeah. And it’s actually becoming more urban. There are protests in the streets of Vancouver right now over a big data center. Seattle just issued a moratorium, which is interesting because Seattle is a big tech hub.

So absolutely this movement is growing, and I think there are different reasons for it. One is people don’t like this infrastructure. It has all sorts of negative consequences. These are absolutely massive build-outs. They often have real consequences for people in the vicinity—from incredible noise that keeps people up at night, that makes people want to move, but by that point it has destroyed their property values.

They’re often run on what should be temporary power sources—gas turbines, methane turbines that have very immediate consequences for the air people are breathing. Sometimes it smells bad, or even if you can’t sense it, there is extra pollution. Depending on the locality, there can be strains on the water supply. They often raise utility bills.

And then people don’t really like what it’s about, right? There used to be a compact, which was, OK, we’re going to do industrial development, but you’re going to get some jobs. Maybe you’ll get a few hundred jobs. You may get a few thousand jobs.

These data centers—sometimes they’re billion-dollar build-outs, and there are 30 jobs, 100 jobs. And in fact, there’s a company now that’s offering robot security dogs to replace the human security workers that were guarding these places.

So the jobs that are permanent tend to be low-wage security jobs or janitorial jobs. The higher-paid work is temporary—it’s in construction or building the actual computers. So it’s a bum deal.

People are also finding out that they are being built with incredible tax incentives that often don’t benefit the community. And so there are all sorts of reasons that people are questioning this.

And then I think there is the bigger context of: but hold on, what does this portend for our collective future? Do we want to live in an AI world? And then amazingly, something as amorphous as AI or the cloud—you go, Oh, it’s actually in my backyard. They’re trying to build it here, and people are realizing that they can fight back.

I’m on some Signal chats with people from 45 states fighting back against these developments and definitely seeing themselves as part of a bigger push. And so I wrote the piece in The Guardian with Saul Levin, who is a longtime environmental organizer—he’s from Michigan, and he’s been on the data center beat for a long time. And we were actually replying to those folks who were like, Oh, is that really the best way to fight AI? It’s kind of whack-a-mole.

And our point is: we’re on incredibly complicated political terrain. It’s actually amazing that there is a space where people can gather, find each other, and push back. And when people do gather, they’re finding out, Oh, actually, we might not have voted the same way. We might not have a lot in common in terms of culture war issues. But hold on, we actually all object to this. And it’s creating these new solidarities. So I think the physical space that these data centers offer is actually providing an incredible opportunity for organizers.

Bacon: So I want to ask, how did this get politicized? And the reason I want to ask this is because it appears the Democrats have decided they oppose them now, but that’s—they followed the consensus.

Taylor: Not all Democrats.

Bacon: Some of them are, yeah. But what I’m getting at is: it seems to me that AI data center proliferation was fine with most elites in media, business, politics, both parties. And yet a groundswell of people started opposing it. So I’m curious—it’s unusual in our culture today. You often find political movements are kind of top-down. Sometimes they’re bottom-up, but usually they’re led by—Black Lives Matter, there were at least active civil rights groups that existed for a long time.

So in some ways, I’m curious: how did these people figure out, Oh, this is something we can oppose? Because a lot of these cities, actually, the city council was trying to hide the tax credits from them. It was not very transparent. A lot of places where the media’s not very strong. So how did people get informed on this?

Taylor: Yeah, that’s a really interesting question. I do want to just linger on your point about the lack of transparency, because I think that’s a huge element that is pissing people off.

Bacon: Oh, that also causes that. Yes.

Taylor: All of these deals are under the cover of these NDAs, where often much of the city council doesn’t even know what’s going on.

Bacon: The government in the city has decided either to cover it up, or they don’t know themselves. The mayor or whoever has done it without them knowing.

Taylor: Yeah. So I did some reporting in Memphis, where Elon Musk has built his Colossus supercomputer, and then neighboring—there’s actually now three of them in Tennessee and Mississippi, all in this area—and just absolute secrecy. And that’s part of what created this incredible........

© New Republic