Concerned Latinos crowding into a rec center to protest a data center in El Paso, Texas. Suburban white homeowners shouting down pro-data center politicians in Festus, Missouri. Black teenagers and elders swamping a local zoning meeting to demand an end to data center construction in DeKalb County, Georgia. Conservative farmers rising up to resist a proposed data center in Howell Township, Michigan.
When many of us look at the grassroots opposition to AI data centers springing up in communities across the country, we see the outlines of the multiracial working class coalition we’ve always dreamed of — a diverse, nonpartisan, top-bottom (versus left-right) movement against Big Tech billionaires that has the potential to reshape American politics in incredibly positive ways.
In her recent Jacobin essay, “Democratic Governance of AI Is the Real Solution,” Holly Buck sees something different. She writes, “What if the picture that emerges of ‘data center resistance’ is one of educated middle-class people … mobilizing, consciously or not, to protect their class position from the threats AI poses?” In this hypothetical, she continues, “It’s not fair for affluent environmentalists and property owners to try to stop development of this infrastructure before most people in the world have even had a chance to work with and learn from these models.”
This style of argument — making an assertion with zero substantiation that flies in the face of existing evidence and functionally carries water for Big Tech billionaires — describes every claim Buck makes to support her piece’s contention that “a moratorium on AI data centers is a terrible idea” that “poses serious equity concerns.” If this is AI boosters’ best pitch to the left — and certainly folks like Matt Yglesias, who’ve been heaping praise on Buck’s essay, seem to think it is — then it is worth responding to her arguments in detail to demonstrate just how fatuous that pitch really is.
But before getting into a point-by-point rebuttal of Buck’s piece, we should be clear: Her argument is not just wrong in the particulars. It is fundamentally incoherent in its basic premise. Buck’s thesis is that a moratorium on data centers “is not a substitute for actual AI governance.” But nobody claimed it would be! The entire thrust of the moratorium effort is to halt the breakneck development of these projects so that we can enact “actual AI governance.” Buck complains that organizations supporting “data center blocking efforts should put their attention towards a broader set of solutions” — but, of course, they have. Organizations like Public Citizen have developed a comprehensive suite of actionable steps to rein in Big Tech and make sure consumers, workers, and the climate are protected from the data center buildout. And they're not alone — other organizations, like the NAACP, have also put out guiding principles for equitable data center development.
The problem is not that we don’t have policy solutions to this crisis. The problem is that Big Tech’s buildout — which is currently designed to serve a few billionaire techno-fascists and nobody else — is moving at a breakneck pace and we don’t have the time or leverage to establish the regulatory framework necessary to make this system work for the public. In the words of Alex Bores, a candidate for Congress whom AI billionaires have spent millions of dollars attacking because of his efforts to enact the “actual AI governance” Buck claims to want, “The Bernie Sanders-AOC proposal is a moratorium until we pass real regulation that protects people. … They are setting the terms of the debate, which is: Why are we going forward with this until we’ve done the real work? I think that’s the right question to ask. If I could wave a magic wand and pass any bill I want, it wouldn’t be the moratorium. It would be the regulations that the moratorium is calling for. But putting that as a negotiating tactic, I think, is meeting the moment in the scale.”
Buck’s piece seems to rest on an assumption that there is a magic wand we could wave to enact “democratic governance of AI.” She writes that “the people should be driving this discussion, not companies like OpenAI.” But how does she propose that the people start driving this discussion? Well, she doesn’t, other than to recommend “public engagement and education on the technology, the stakes, and the policy options.”
Let’s be clear: Big Tech billionaires are putting hundreds of millions of dollars into super PACs to block any attempts at “actual AI governance.” Public engagement and education is a laughably meaningless answer to this political leviathan. The passage of data center moratoria is — here in the real world, where we don’t have magic wands — essentially the only tool available to us that could exert meaningful leverage over these companies, as is unquestionably necessary to enact regulation of their dangerous practices. For Buck to fail to engage or even acknowledge this dynamic is, by itself, more than fair grounds to dismiss her argument as unserious.
Of course, Buck does not just limit herself to this absurd premise. She supports her premise with a series of equally unsubstantiated supporting arguments.
Her first assertion is that Big Tech’s response to a moratorium would be to simply offshore their data centers. This is ridiculous on two fronts. First, developing countries do not have the energy and grid infrastructure to support these projects, absent absolutely massive investments by Big Tech. What’s more, offshore hyperscalers would respond to queries significantly more slowly than domestic ones, creating another huge problem for AI companies. Big Tech will go to great lengths to avoid these costs — hence the leverage discussed above.
Second, at a more conceptual level, Buck’s argument — that we should be “wary of proposals that would send burdens elsewhere” — is tantamount to saying, “You must let the ruling class hurt you, because otherwise they might hurt someone less privileged than you.” This is essentially an argument against any jurisdiction in the U.S. passing any pro-social regulation at all, ever — an extremely convenient assertion for the forces of capital, but an utterly ridiculous one for those of us who want to improve the lives of real people in the real world.
Buck’s next major contention is that the moratorium push is a “dead end” because the “disparate right-left coalitions that have emerged around stopping data centers have different interests when it comes to other issues.” This is a sentence that could only have been written by someone who has never engaged in on-the-ground organizing or political work at any point in their lives (yet still, somehow, feels empowered to opine on that work to the people who are doing it, day in and day out). Building coalitions among constituencies with “different interests when it comes to other issues” is quite literally the entire work of politics, and always has been! As @Ben_Inskeep, an organizer who’s been supporting these fights in Indiana, put it: “People from all walks of life are waking up to the broken system we have where both parties are exploiting us. Not a ‘dead end.’ An incredible opportunity for class consciousness!”
After dismissing the basic work of political organizing, Bucks next attempts to minimize the climate impacts of Big Tech’s data center buildout, arguing that while “the emissions and energy draw are real issues,” they “need to be placed within the bigger picture of our climate challenge” — i.e., they’re only a portion of the emissions we produce overall (yes, that’s literally her argument). Of course, Buck ignores the reality of how data centers’ energy demand is operating to actively extend the life of dirty coal plants and drive a massive build-out of new gas-fired power plants. Buck then repeats her original flawed premise, saying that this problem is “something that states can regulate.” Yes, it is — and passing clean energy requirements for data centers is one of those policy solutions that community groups are advocating for, and that a moratorium could help local and state jurisdictions actually establish.
Finally, in order to frame AI access as an equity issue, Buck makes some grandiose claims about the liberatory potential of AI use. Her evidence offered on this point is entirely personal — Buck describes how she came to believe, after taking several undergraduate courses last year at the school where she works at a professor, that “for many subjects, the personalized tutoring offered by AI is far better than the outdated lecture-based model still employed by universities.” God forbid people have to live in a world where we haven’t replaced all our professors with AI! It’s quite the admission from a professor — and one that makes me extremely grateful to not be a student stuck in one of Buck’s classes.
This is not to write off the potential value AI can provide to an individual (though it’s worth noting that according to Buck’s immiseration-by-AI-deprivation taxonomy, I — a lawyer and graduate of elite universities, where I was taught by professors rather than digital tutors — am a member of this new underclass because I refuse to use AI in my daily life). But this over-emphasis on AI’s empowerment of the individual obscures the reality that the benefits of AI overwhelmingly accrue to the corporations deploying them, not to regular people, and that — as currently designed — AI works to entrench and exacerbate existing power dynamics. From documented cases of algorithmic discrimination in biometric systems like facial recognition, to disastrous experiments in integrating AI chatbots into public school systems, to the use of AI to justify damaging benefit cuts and Medicaid coverage denials to low-income people, to RealPage’s algorithmic facilitation of rental price-fixing, to Amazon’s AI-supported surveillance of warehouse workers, to the increasingly prevalent algorithmic wage exploitation of gig workers, to the growth of the dystopian state surveillance of every American, to the fact that I can’t ever find any goddamned customer service phone numbers anymore because they’ve all been replaced by AI chatbots, the actual evidence is clear that AI is not liberating us as individuals — it’s concentrating power in the malignant hands of those who already have far too much control over our lives.
I don’t know Buck, and cannot comment on what motivated her to make a series of arguments so divorced from our reality. But I can describe what these arguments do on a functional level — whether that was their intended effect or not — and that is to carry water for an industry that is growing increasingly concerned its toxic public image could affect its ability to continue vacuuming up our land, water, and energy in order to further enrich its CEOs. A genuine working class coalition is forming in opposition to this greed and avarice. That movement deserves to be celebrated and supported, not denigrated and lied about. And we deserve a future we control, not one dictated to us by Big Tech billionaires and their apologists.