AI won't destroy us (if we destroy ourselves first)

Over the last year, the small city in upstate New York that I call home has become an unexpected hub of organizing against Flock cameras. As soon as the presence of these so-called "Automated License Plate Reader" systems were discovered by a local resident—one of more than 25 of the AI-powered devices were installed throughout town without public consent or knowledge—activists swiftly became a thorn in the side of the city's right wing mayoral administration. That mayor went as far as declaring a state of emergency in order to protect a $156,000 contract renewal payment to the Georgia-based surveillance company. The story has been covered in national papers, making our little city a known quantity in the wider "DeFlock" movement.

Having been very lightly involved in this effort—mostly offering a small amount of communications support—I am also privy to several local anti-Flock chat groups. It was in those chats that I recently witnessed an episode which threatened the future of the local effort. Insults were thrown, complex and time-sensitive projects were derailed or disrupted, contributors departed, social media accounts were seized and others locked out, arbitrators were involved.

The row wasn't the result of any major ideological dispute, breach of decorum, or accusations of abuse. Rather, the breakdown occurred when an affiliated group was accused of possibly having used generative AI to create a minor graphic element for a digital event flier. The most vocal protest came from a leading member of the coalition, who is also a visual artist; as such, they were outraged over the perceived betrayal of values and solidarity, the use of something that threatens their creative profession, and even society itself.

Based on subsequent conversations and evidence, all detailed and debated exhaustively by the group, it appears that the image was not actually made with generative AI. Honestly though, even if the flier had been made with AI, so what? To my mind, the whole episode exemplified a counterproductive, reflexive posture that often emerges in the face of emerging technologies, and which may stand in the way of a better—or at least more benign—arc of development for AI.

It has become something of a fashion to oppose and avoid contact with anything even tangentially related to generative AI. Often these are particularly intelligent, sensitive, principled people—good people with good intentions and sound critiques. Increasingly, though, to remain consistent in this refusal to engage requires an eschewing of many other technologies that underpin daily life.

We can soon expect AI to 'enhance' anything with an electronic pulse—our refrigerators, cars, phones, crib cameras, pacemakers, bidets—regardless of our feelings on the matter. I really don't want Google Gemini parsing my emails—I certainly didn't ask it to, and yet it does. Meanwhile, and more consequentially, AI is being deployed across the economy, gradually becoming unpredictable, black-boxed infrastructure for innumerable industries, even governments and militaries.

All of this is underpinned by the environmentally ruinous proliferation of resource-chewing, smoke-spewing data centers, with untold collateral damage done to to planetary and societal health—and spurring a forceful resistance movement that may prove invaluable to broader anti-oligarchic organizing. Like most powerful new technologies, the social and economic consequences of AI are largely unknown, though we have some ideas of where it may be headed. Mostly, the prognosis is not very good—the top leaders in the field openly anticipate their works may lead to the eradication of the human species—while the rosiest predictions are not very credible.

The alarmism and calls from the industry for government regulation may actually be part of a marketing push, and an effort to pull up the proverbial ladder up in order to prevent new players from entering the arena. Not for nothing, the same people warning of existential civilizational risk are also describing their products as literal genies that can “grant any wish," so we can probably take both poles with a heaping of salt.

The forward march of AI is frequently described as inevitable. Yet the fact remains that this technology, like every other, is neither developing nor deploying itself. Its alarmingly rapid advance is less a function of the technology's inherent traits than the voracious appetites of the capital interests behind it. Should we be surprised that a technology which promises to replace vast swathes of human labor finds its biggest champions in capitalists?

Hopefully it is clear that I am not among the barking seals who see in AI a solution to all of humanity's problems. My own position is that the technology is fascinating and has some very promising applications, all of them far more narrow and limited than would justify even a fraction of the blitzkrieg, ravenous development of the technology and its physical infrastructure. I, like my more decisively critical friends, find the slop and the waste and the mental health impacts and the copyright infringement and the attendant laziness and cognitive atrophy and all the rest of it to be disturbing and bad. It will probably prove to be a huge problem that it is now possible to reverse engineer essentially any piece of software.

I also don't think these negative outcomes are inherent to the technology, nor is checking out completely from its development and implementation a hedge against those outcomes. In the same way that stomping spotted lanternflies on sight isn't a solution to their proliferation, the particulars of AI's drawbacks—copyright infringement, resource consumption, endless fonts of brain-rotting content—are not the root of the problem.

I once heard technologist Liz Howard articulate that most of the righteous and valid criticism of AI amount to critiques of capitalism. This strikes me as deeply true. Technologies emerging since the industrial revolution tend to follow what is called a hype cycle—a flash of enthusiasm and possibility leads to explosive, experimental growth, followed by a hangover as the promise is never quite realized, and unexpected consequences emerge. Then, with time, adoption gradually increases and a more balanced relationship with the technology emerges. Calling this a cycle is a bit misleading though, because it is all based on contemporary concepts of innovation, which themselves are built on the linear, unsustainable extraction of resources. In short, the development of a technology is shaped by the profit-oriented proclivities of capital, and as technologies become more powerful or attractive, the outcomes are profound. Whenever I drive on a road among in and among other cars, I feel as if I’m living in a post-apocalyptic world in which the machines have already won.

AI is here. There is little use in pretending otherwise, nor that it can be unwound or uninvented, stuffed back into Pandora’s box. Famously, that’s not how the box works. This doesn’t mean the alternative is to embrace and promote this technology and all of its attendant bills, and it doesn’t mean that artist should be OK with seeing people resort to resource intensive alternatives to simply paying an artist for their work. But neither should the predictable cycle of development deployment, boom, and bust that accompanies any powerful new technology be a surprise, let alone a reason to disengage entirely from the arc of its progress.

People are curious; people innovate; people have to know the ultimate potential of every tool they invent. The atomic bomb did not emerge from the pressures of war so much as the possibility that it might exist or be technically possible. This is why, rather than a rejection of the technology itself, I hope that we can instead cultivate a culture that insists on powerful new tools are developed and deployed according to something other than a profit logic. Amidst the boom and bust of the markets—and, increasingly, supremacist and pseudo religious ideologies of the sort that often fester in over capitalized corners of society—that define 'tech development', there's an imperative to forge alternative relationships to, conceptions of, and applications for AI.

Rejecting that opportunity—and dare I say responsibility—leaves it to the worst actors to define a technology that could be put to many good and interesting ends. For example, Abi Awomosu has called for a counterculture movement in AI akin to punk; the formation of subversive, decentralized layers of discourse, distribution, and culture that allow for a technology to grow in ways that capital ownership wouldn't consider or permit—the ways that may in fact empower the broader public.

AI may be the clearest example yet of Paul Goodman's dictum that technology is far more the province of moral philosophy than science. It is disruptive in the familiar sense of the word, though 'disastrous' may be a more appropriate term. The AI industry—as distinct from the technology of large language models itself—is already rotten with problems and risks. Mental health, the ability of children to socialize, consensus reality, privacy, livelihoods, the entire economy—all and more are at stake. There is little reason to also risk the dissolution of the very communities resisting unchecked development and deployment of AI technologies, data centers, and surveillance. Such groups are, after all, the only hope for achieving an outcome other than what's on offer from industries and their owners who are all-too-eager for omnipotence.