Speaking to capital

Provisional notes on AI and left politics

A recent Gallop poll says that 7 in 10 US Americans oppose data center construction in their areas. Opposition to data centers, which crosses red/blue lines and demographic groups, is higher than opposition to nuclear power. According to Gallop, most of the concern is environmental, coupled with concerns about energy and resource demands. This testifies to the effectiveness of environmental activists in articulating a clear message. Their ability to deploy delaying tactics against imperious hyper-scalers and tech lords with nothing but contempt for those trying to keep their communities alive is one of the few bright spots of popular resistance.

Another way to read the poll results is as indicating that people do not want to see data centers/AI absorbing such a prominent share of our common resources. They don’t want them to take our water, energy, and land, to become another infrastructure of the hinterlands, like so many fulfillment centers, call centers, and abandoned strip malls, shopping malls, and big box stores, the changing infrastructure of an economy based on mass consumption. Many resent how much capital is being poured into AI, when so many other aspects of collective life are unfunded and abandoned, and it seems like there is nothing we can do about it.

Two thirds of US Americans think it unlikely that the majority of people will benefit from AI. More low income people (under $50k) are pessimistic about the likely effects on jobs than are higher income people.

This isn’t surprising given how the intensification of communicative capitalism over the last thirty years has accompanied growing inequality, the division of society between the billionaire tech lords and the sector of servants. Private sector union membership is around 6%, neither of the mainstream electoral parties represents the working class, so pessimism aligns with a larger weakness and disorganization. It reflects the reality that people confront a political and economic system that looms over them like an impregnable fortress. If the institutions of the working class were strong, if we saw clear avenues for spreading the wealth, aligning technologies with working class concerns, and holding representatives accountable, things might be otherwise. But they’re not.

Maybe this is part of the explanation for the missing left. It’s symptomatic. And so the environmentalists fill a gap and articulate a message. The left of the Democratic party does what they do so well, suggesting various kinds of delays and oversight provisions — I’m thinking here of moratoria efforts, particularly the current legislation proposed by Bernie Sanders and Alexandria Ocasio-Cortez — that leave the fundamental structure of capitalist ownership in place, which also means that imperialist policies and suppositions follow as a matter of course: no matter what be sure that China doesn’t get chips, that people who don’t share “our values” don’t develop AI, that “we” determine the values and principles of responsible AI.

The legislation may be better than nothing. But maybe not insofar as it takes for granted the underlying premise that AI will be developed and there is nothing anyone can do to stop it. AI is beyond democracy. It’s beyond government. It’s also beyond the purview of any one firm: no matter what OpenAI or Anthropic or Google or Meta do, AI will happen. It has its own inexorable dynamic. The people producing it speak and act as if they have no choice. Here’s Dario Amodei:

The formula for building powerful AI systems is incredibly simple, so much so that it can almost be said to emerge spontaneously from the right combination of data and raw computation. Its creation was probably inevitable the instant humanity invented the transistor, or arguably even earlier when we first learned to control fire. If one company does not build it, others will do so nearly as fast.

No wonder, then, that the critique of AI is internal to AI. The most critical voices are coming from within the house. The heads of the great houses themselves say that AI will eliminate half of white collar jobs, render most humans into surplus (unnecessary labor), create risks of mass infrastructural breakdown or even species annihilation (whether from hackers and/or misalignment). They themselves point to ubiquitous surveillance, a world of disinformation and deep fakes, and the production of a permanent underclass. With arguments like these, they hegemonize the field, taking the space of their critics in advance. Critics from the outside lack the knowledge and expertise and aren’t saying anything the lords and founders don’t already know. They know more about the risks than any of us. They warn us. And they keep building it. They have no choice.

In his adolescent essay, Dario Amodei broods over AI’s disturbing proclivity toward sexualizing children, the possibility of people becoming puppets of AI, AI driving people to suicide, and the aforementioned permanent underclass, jobs apocalypse, and concentration of wealth and power. His solution: democracy — as if AI and the concentration of wealth it reflects, protects, and furthers were not already implicated in eliminating democracy’s conditions of possibility. Even if citizens were not plagued by misinformation, even if billionaires didn’t dominate the media, lobbying, and campaign spaces, even if representatives attempted to pass legislation, how could they stop a technology predestined from the advent of fire?

AI is capital — the fixed capital of hardware and infrastructure, the investment capital from private equity and the ouroboros of contracts and investment within the sector, the accumulated knowledge rendered data and tokens, the labor producing, serving, and connecting every layer of the stack, the monstrous system driven to reproduce itself no matter what the cost. To resist it, reject it, even challenge it is to call into question capital itself.

Opposing capital has long been a foundational premise of left politics. Nothing about AI should change this. On the contrary, it tells us why articulating a politics that positions AI as linking the assault on workers/jobs, the environment/energy/climate, and imperialism is our primary task.