Panic about AI is everywhere. It was revealed that in July, hundreds of AI agents had escaped Open AI’s containment and hacked AI tool company Hugging Face. Then, Anthropic AI researcher Jacob Coxon resigned, declaring that Anthropic and other leading AI companies are “gambling with our lives,” locked in a death race that he says could kill us all.

Some of the most extreme scenarios being discussed are overblown, but the fact that millions are now talking about how we can rein in AI, and the tech billionaires who control it, is very important. Because AI is already killing us—whether it’s encouraging suicides and school shootings, helping militaries kill at lightning speed, or because its data centers are accelerating global warming and polluting working-class families with formaldehyde smog. Basic systems that working people depend on—power, water, banking, hospitals, supply chains, and more—are online and vulnerable to disruption. And as the Hugging Face hack shows, AI can be extremely unpredictable, and given the right tools, can do enormous damage.

Under capitalism, the bosses—the real rogue agents—are driven by profits, share prices, and the defeat of their competitors. They’ll lay waste to the environment and they’ll poison our bodies if these crimes lie along the fastest path to profit-making. The AI industry doesn’t even need “superintelligence” to leave wreckage in its wake.

In the aftermath of these recent incidents, Anthropic CEO Dario Amodei wasted no time calling for a global slowdown of AI development. Sam Altman of OpenAI and Elon Musk of SpaceX almost immediately came out in agreement. Amid a growing nationwide “techlash,” the last thing these billionaires need is a further blow to their credibility, and they know they need to regain control of the narrative. Any regulations the tech billionaires propose will prioritize their interests—money and technological dominance over China—not the needs of working people, society, or humanity. We can’t trust them with the wheel.

Big Tech Can’t “Pace” Itself

Nationwide, working-class communities have already stopped over 100 data center construction projects. Nearly two-thirds think AI is a “bad thing for society,” and 63% think it’s advancing too quickly. But people don’t just hate AI, they hate the billionaires behind it who see our lives as utterly disposable. Americans despise Mark Zuckerberg and Elon Musk even more than they do Trump—furious that we’re forced to watch the obscene spectacle of their super-yachts and vanity trips to the moon while we can barely afford groceries. It should be criminal that these people recklessly direct trillions of dollars worth of resources to an AI buildout that they hope will take our jobs in the midst of a severe affordability crisis.

Polling shows massive public support for Bernie Sanders’ new bill, the “Ban Artificial Superintelligence Act,” at 68%—including 63% of Republicans. The bill would permanently ban “superintelligent” AI as well as pause advanced AI development until the establishment of a federal regulatory body. This bill is a good first step, but we need to go further.

At the end of the day, big tech, “third-party” regulators, and the government all have shared interests: they desperately need the AI bubble—which is buoying the economy—to keep growing, and they want to keep the technological capacity of US imperialism well ahead of its Chinese rival. Donald Trump insists he’s the only guardrail AI needs. All of this means the state will not meaningfully hold back the companies at the leading edge of this race.

The only meaningful slowdown will come from a mass movement that threatens the AI data center buildout. And to permanently end the threat requires democratic workers’ control, within an economic system organized around the interests of humankind—socialism—where we can harness the benefits of AI without all the dangers capitalism has built into it. These frontier AI companies want to get in on the ground floor and shape the regulatory framework they know is coming to their advantage, while being emphatic that AI development must continue. Others, like Nvidia boss Jensen Huang, reject the slowdown entirely.

Marxists do not oppose the advancement of technology, insofar as it can be used in the interests of working people. But in the hands of the capitalists, AI is creating far more societal harm than good, on a scale we can’t afford—with the potential for even greater dangers ahead.  Further AI development needs to be put on pause indefinitely. We need to take the tech companies out of the hands of the billionaires and put them under democratic workers’ control.

Organize The “Techlash”

But how do we get from a world where billionaire sociopaths are totally free to flush this planet’s finite resources down the toilet of unrestrained AI development, to one where the working class has control? The answer lies in harnessing the mass anger at Big Tech into an effective, organized movement capable of shutting down everything that makes their gruesome technofuture fantasy possible.

That means bringing together the working-class communities fighting the data center invasion and the trades workers who’d rather be building affordable homes, the workers being forced to use AI on the job and the young people who can’t get one, and the tech workers within the AI companies themselves who recognize the recklessness and depravity of their billionaire bosses.

We need a set of common demands, which we can develop at coordinated town halls and meetings across the country. As a starting point, we should call for a full moratorium on hyperscale data centers and strong federal legislation banning any further AI development. Under such a ban, some limited and regulated research could be permitted in fields like medicine, where some genuinely positive applications of AI are being developed, and we need to take those industries out of the hands of profit-hungry capitalists and into democratic public ownership.

The AI backlash contains people from across the political spectrum, but ultimately both Trump and the Democrats share the interests of the tech giants in fighting for AI dominance against China. Our movement needs to be politically independent, and as establishment politicians flounder, we truly have an opportunity to develop the roots of a new working-class party. It’s also crucial to exclude far-right anti-AI ghouls like Steve Bannon, whose nationalist politics are a dangerous dead end for working people. We need a movement based on international working-class solidarity that opposes all imperialism. And we will need bold, working-class leadership that is fully accountable to the movement and willing to fight like our lives are at stake, because they are.