The AI Backlash Is Transforming Into a Political Movement
· The Atlantic
If there’s one thing America seems to agree on, it’s AI. Some three-quarters of Americans oppose a data center being built in their area, three-quarters of Americans think AI will hurt their job security, nearly 70 percent think the technology is advancing too quickly—the polls go on. Progressive and Republican candidates alike are bashing the tech industry to win primary elections. States and cities are halting plans for data-center development.
Visit asg-reflektory.pl for more information.
Yet in cloistered San Francisco, where I live, staffers at OpenAI and Anthropic are developing this same feared technology at light speed and pulling down compensation packages that stretch to considerably more than $1 million. Silicon Valley’s growth is largely unencumbered by regulation, and President Trump’s administration has if anything used its authority to speed the AI industry along. The divide between these camps—the majority of people in this country who simply have to deal with the technology as it’s rolled out, and the much smaller population of people with direct influence over it—is a defining feature of this era. And so I noticed when a small group, with ties to both the Bay Area and Washington, started making noise, insisting that the AI race slow down and the technology be overseen not by executives and engineers, but by the broadest population possible.
The group—which formally launched today and calls itself Irreplaceable, as in people are irreplaceable by machines—has an audacious and perhaps naive vision. It is essentially a protest against capitalism as it is practiced in America today, focused on resisting first and developing a coherent plan for the future later. I have seen exactly how bulletproof the AI companies are—pilfering seemingly infinite volumes of media to train their models, building unpermitted power plants for their data centers, and unleashing autonomous agents capable of massive security breaches, almost entirely without consequence. I wanted to meet with Irreplaceable’s founders to ask: How do they possibly expect to pull this off?
[Read: It may be time to panic about AI]
The organization is staffed with people who made a name for themselves in the climate movement, and they are hoping to avoid past mistakes by focusing on the emotional reaction Americans are having to AI, rather than anchoring their work in white papers and policy recommendations. “People get it,” Jeremy Ornstein—a campaign strategist for Irreplaceable and former organizer with the Sunrise Movement—told me. “People have the intuition that if we create machines more powerful than people, that could go dangerously and terribly wrong.”
Phil Aroneanu, Irreplaceable’s executive director and a co-founder of the climate organization 350.org, told me he believes that climate advocates initially floundered because they acted as policy wonks. They thought that making evidence-based arguments about the dangers of melting ice sheets and rising sea levels to receptive congresspeople would be enough to pass nationwide climate legislation. Today, Aroneanu said, movements are built not necessarily on what people think, but instead on what they feel. For climate, that meant fossil-fuel-divestment campaigns, protesting oil pipelines, and the school strikes led by Greta Thunberg. People are already “feeling the squeeze,” Aroneanu told me. “And we should be pointing that anxiety and that anger in the right direction.”
Irreplaceable will prioritize a range of issues: mitigating automation’s effects on the economy, protecting against government and corporate use of AI to surveil Americans, and establishing industry oversight and guardrails to prevent catastrophic disasters, such as widespread blackouts caused by AI-powered hacking or bot-engineered pandemics. As I spoke with the activists, I kept thinking about what had become of the climate movement—the Green New Deal passed, but only after being watered down, and it was then gutted by the second Trump administration—but their populist approach makes sense. Some of the biggest and perhaps only lasting successes over the past 15 years of American politics—the Tea Party, Alexandria Ocasio-Cortez, MAGA, and election denialism—have been fueled by outrage over an unfair economy and seemingly rigged politics.
These organizers have demonstrated their ability to win significant battles. Sunrise, 350.org, and others were instrumental in passing national legislation, galvanizing widespread climate demonstrations, and accelerating private-sector development of renewable energy. And Irreplacable’s members have already flexed some muscle when it comes to AI. In July, hundreds of people gathered for what was likely the largest anti-AI demonstration in San Francisco’s history, marching from OpenAI’s headquarters to Anthropic’s and finally to Google’s. A few months earlier, some of these “Stop the AI Race” organizers had staged a somewhat disorganized protest: Attendees at times struggled to chant in unison. The second time around, the march was advised by Ornstein, and attendees had chant sheets, a marching band, and a slushy truck (to “freeze” AI).
Already, a number of climate-oriented groups (including the environmental-justice arm of the NAACP as well as the Sierra Club and Earthjustice) are supporting local data-center battles by filing lawsuits, shaping model legislation, and organizing protests. The overlap is both concrete—data centers pollute the air and burn fossil fuels for their tremendous electricity demands—and thematic: a movement against corporate interests to change or regulate the technology of the future. At least on a local level, the groups are finding victories: Dozens of data centers have been canceled in the face of local opposition, and all of the major AI companies have promised to pay extra to avoid hiking up local electricity bills. In Michigan, both Abdul El-Sayed and William Lawrence, who is a co-founder of Sunrise, won Democratic primaries as progressive candidates hammering against data centers.
Still, this bubbling AI anxiety is far from a sweeping tide. That growing numbers of Americans are concerned about AI development does not mean they are concerned for similar or even compatible reasons. People may say they loathe data centers, but they might also enjoy the conveniences of AI products at home or work. They fear bots getting too powerful, too quickly, but also think America needs to stay ahead of China in AI development. At one anti-AI protest, I saw people who were ostensibly on the same side disagree about AI’s water use and whether certain jobs would be better off automated by chatbots. And to the extent that a sufficiently powerful AI movement would need to be global, just as climate organizing was, the technology generally polls much better outside the U.S. The AI backlash is to some extent an American phenomenon, which bodes poorly for any sort of global norms or treaties.
[From the April 2026 issue: Inside the dirty, dystopian world of AI data centers]
When pressed, nobody from Irreplaceable and no other climate activists turning to AI would articulate a concrete AI agenda. “We’ve aligned around saying no,” Lawrence recently told me. “Now, there is an important conversation that is yet to come, about—okay, so what do we want?” This is, for now, a significant weakness; a key turning point for the climate movement was providing a positive vision, the Green New Deal, for both decarbonizing and reinvigorating the American economy. Being able to link every data-center battle and primary campaign to a national vision would give the AI backlash some form.
It’s not wrong to want to develop a policy road map with the input of a broad coalition; climate advocates took decades to devise and unite around the Green New Deal. Irreplaceable’s stated plan is to build a coalition this year, develop a unified policy next year, and implement that agenda by the end of 2029. Measured against any metric other than Silicon Valley’s warp speed, that would be fast. But those few years will leave plenty of time for tech companies to push back—they’ve already spent tens of millions of dollars leading up to the midterms—and regardless, by then OpenAI, Anthropic, and any number of Chinese firms may already have built and lost control of even more powerful models.
Irreplaceable is wagering that people can move faster than capital. Absent specific policies, they are basically placing a big bet on populism—that the issue is not any particular chatbot or data center but the companies building them. The coalition comes first; the policy, after, Aroneanu told me. “You have to show the benefits to people immediately. It’s gotta land in their pocketbooks almost immediately for it to have political salience. And then you have to keep defending it.” Opinions about specific aspects of AI may differ, but fundamentally, people don’t like that “this technology is coming into their lives and they have very little control over it,” says Caroline Behringer, a communications strategist who worked on President Obamaʼs Climate Action Plan and now runs Athena Insights, a polling organization that studies attitudes toward AI. The movement, in Aroneanu’s words, must be polarized not left-right, but “down-up.”
Of course, Irreplaceable’s climate roots link it to progressive causes, and the group will primarily focus on organizing the center-left around AI. When I asked Aroneanu if he worried this might be counterproductive—polarizing the issue and allowing for the kind of partisan rancor and backsliding that beset climate activism—his response was pragmatic. “Chances are, this issue becomes polarized,” Aroneanu said. “I think that political parties are really exceptionally good at generating negative partisanship,” just as industry is likely to “split the base.” At some point, the climate movement had to decide that the best strategy was to accept partisanship and build power from the left. Although Irreplaceable hopes that conservative counterparts will emerge and believes its populist message will cross culture-war lines, Aroneanu does not see this as his explicit remit.
Broad, at times even ambiguous, backlash may make sense if AI is a general-purpose technology, as Silicon Valley is fond of saying—which is to say, if the technology touches and even upends nearly every aspect of the economy. Creating a technical oversight body for AI cyberattacks will not protect every American’s livelihood during mass layoffs; wealth redistribution will not prevent AI companies from spying on your most intimate interactions with a chatbot; voting won’t matter as much if AI companies become more powerful than Congress. Major figures at both OpenAI and Anthropic have suggested that a true AI revolution could require new democratic and economic institutions. Maybe they’re right. And maybe these companies, which have already torn down so much, should not also be the ones in charge of rebuilding.