One of the just-so stories we keep hearing about AI is that it’s inevitable, that the technology is here and will continue to be here, and we better get on board or get left behind. These stories have the ring of a threat because they are, explicitly and otherwise, threatening. They are also familiar.

Fear that there may be no alternative to the will of the AI arise because we have been told for decades that there is no alternative to neoliberalism, that there is no alternative to the mediation of all society by profit-driven markets, no alternative to the universal power of private self-interest that continually tries not to better the world, but to maximize it’s own profit and hence power. Stories about the “promises and perils” of AI ring true, not because the AI is poised to hunt all of us down, but because the stories reflect real experiences of technology, capitalism, and ideology; they reflect the capitalist developments of the incomprehensibility of technology, the invisibilization of labor, enclosures, proliferating neoliberal bureaucracies, and the sense that there is no alternative to capitalism and the status quo.

Blix & Glimmer, Why We Fear AI, page 56

In other words, the threat isn’t so much that AI is inevitable as that the ongoing—and likely expanding—immiseration of workers is unstoppable. This is the subtext of the strange and conflicted messaging that we get from the hype men: when they say that you better learn AI or be left behind, they are admitting that a great many people will be left behind. And if you—smart and clever and hardworking person that you are—are somehow able to make it to the other side of the line, you’re supposed to find relief or pride at having done so, and not horror at all the people suffering in your wake. You’re supposed to be as uncaring as the capital that uses you.

But getting through this gauntlet is no guarantee of getting through the next one—and there will be a next one, because the plain aim of the technocrats is to immiserate everyone, eventually. From the capitalist perspective, anyone with skills enough to negotiate a comfortable wage is a cost in need of cutting. Add to that the fact that AI’s whole pitch is that the more you use it, the more data it gathers, the more likely it becomes capable of mimicking you well enough to convince the fools above you that it can do your job. So get-in-or-get-left-behind is something of a trick—everyone is left behind, eventually.

Which is both terrifying and clarifying. Terrifying in that the capitalists really do have the ability to do us harm—they have been doing so, already. Clarifying in that there really isn’t any reason to stay on the path they’ve laid out for us. It leads nowhere good. Meanwhile, there aren’t very many people up ahead, and there are a whole lot of us back here. Let’s see what we can do.

I’ve talked about one just-so story of AI—the notion of its inevitability—and I want to talk about another, that AI will increase productivity. This is a somewhat tricky story to explore, because it rests on the obfuscation of what we mean when we say “productivity.”

[C]ertainly, companies will be interested in tracking their customers with AI, whether as targets of ads or as imagined thieves. For most companies, however, there is an even bigger target at which to point their AI technologies: the people they employ. When companies do so, the ostensible purpose is usually simply described as increasing productivity. After all, who could be opposed to getting more done with less work? Alas, increases in productivity are deeply interwoven with two other purposes: first, the automation of supervision and control—management. Second, the reduction of wages, for instance by increasing the pool of workers that can be hired for particular tasks—deskilling, outsourcing, and globalization.

Blix & Glimmer, Why We Fear AI, page 79

It’s worth teasing something apart here. When workers talk about increasing their productivity, they often speak of getting more work done in the same amount of time. As they develop skill, or as the work becomes more automated or more regular, they are able to do more of it. But when companies talk about productivity, they are much more likely to be talking about the cost of the work. The descriptions are, at some level, equivalent, but they emerge from very different political standpoints and have entirely different impacts on people’s lives.

For example, the automation of management improves productivity by reducing the number of managers needed to keep work moving, at times even down to zero. As Blix and Glimmer note, Amazon warehouse workers may find their experience of management is entirely subsumed under automated video surveillance in which there is little human oversight—or, in which the human oversight is itself automated and distant. But we see the same automation drive in more so-called professional labor, too, e.g., when a software engineer is evaluated on the number of pull requests they submit, or a doctor is measured by the change in blood pressure of their patients. Both moves replace human judgement with a purportedly objective system that can do the work of supervision without a supervisor. If when you look for productivity increases you’re only looking for people doing more work, you may miss the fact that a lot of those people are no longer around.

Likewise, we are wont to assume that deskilling looks like someone doing more menial work after most of their work has been automated away. The copywriter who once generated sentences and ideas from their brilliant, creative mind but is now tasked with babysitting a sycophantic LLM that spits out uncanny but plausible-sounding versions of the same is an obvious victim. But deskilling shows up in other ways, too: the copywriter who retains their job as an actual writer—because such work remains valuable underneath an avalanche of slop—finds that they are pressured to do more and more work, at lower and lower wages, because there are legions of other people who can do it, too. The deskilling occurs at the level of the community, not only the individual.

In other words, a synonym for “increase in productivity” is “fewer workers.”

This is the real-life version of the industry fable of the so-called “10x engineer,” a mythical engineer that allegedly adds 10 times the value of a normal one to a company, and the real mechanism behind it: the value that is “added” is literally the wages that the other nine workers are no longer paid, and which thus remain on the credit side of the company’s ledger.

Blix & Glimmer, Why We Fear AI, page 107

This puts those now-ubiquitous AI mandates in a slightly different light: if every engineer (or copywriter, or doctor, etc.) is required to develop the skill to use these tools, then all of them are eminently replaceable by anyone else. So long as lots of other potential 10x workers are waiting in the wings, there will be downward pressure on wages for that lone, last-standing worker. Maybe, they think, if they work really hard and become that mythical 10x engineer, surrounded by an army of obsequious agents bent to their will, they will succeed in earning all of the wages of the nine workers they replaced. But ten times zero is still zero.