Articles & Videos

Despite worries about the investment boom, there are signals that global demand for AI computing remains robust.

Elon Musk is officially the richest man on Earth. The source of his wealth is not his own cunning. Nor can it be fully explained by capitalism’s financial architecture, tax loopholes, and the value created by the workers employed in his various ventures. Tech billionaires’ companies are robbing humanity of its distinctive capacity to create. [...]

With the artificial intelligence boom reshaping the global economy, workers are racing to defend their rights as the ground shifts beneath them. A narrowly averted strike at Samsung Electronics in Korea this spring was a major test of labor’s ability to fight back. Weeks before the planned strike, a historic 40,000 Samsung workers rallied outside the company’s [...]

Instead of schools trying to replace teachers with AI, teachers could use AI to become better educators.

Robert Siebelink was staring down the kind of ballot California specializes in: 61 people running for governor, and that’s just the top line. So the 54-year-old Democrat from California did what a growing number of Americans are doing, according to a story Jennifer Medina wrote in The New York Times on July 4. He pulled up Claude, Anthropic’s chatbot, uploaded his ballot, and asked which candidates fit his values. It helped him narrow the governor’s race down to two Democrats and talked through the strategy with him. He finished the whole thing in a half hour.And it seems like he had plenty of company. A woman in Los Angeles County photographed her ballot and flat-out asked Claude who to vote for. A man in Baltimore told the Times that researching his last ballot ate up something like 20 hours of his life; with Claude summarizing every candidate for him, this one took an hour. Medina’s read is that 2026 might be the first cycle where enough voters do this for it to matter, and honestly, that feels conservative to me.Now … the part the Times didn’t get into.Three days before that story ran, the Federal Trade Commission proposed a policy declaring that AI companies that steer their chatbots toward “undisclosed ideological objectives” may be committing consumer fraud (the public can comment through July 31). Which sounds reasonable! Nobody wants a secretly ideological chatbot. But then there’s the obvious follow-up: Who decides what “ideological” means? Right now, that would be the Trump administration. The same administration that has spent the past year attacking AI companies as woke, that cut the entire federal government off from the one AI company that told the Pentagon no, and that handed the cheapest deal in its AI purchasing program to the company whose chatbot had spent the better part of a day praising Hitler a couple of months earlier.Americans started trusting chatbots with their ballots at the exact moment the federal government finished building the machinery to control what those chatbots say. I think that’s a pretty big story! The right ran this pressure campaign against newspapers for decades, then against Facebook and Twitter. The chatbots are the next phase in a familiar playbook.Back during the 1992 campaign, Republican Party Chair Rich Bond explained to The Washington Post why the right complained so relentlessly about the “liberal media,” and his answer was disarmingly honest: It’s the same thing coaches do to officials, where “what they try to do is ‘work the refs’” in hopes of friendlier calls later. Complain loudly enough, long enough, and the calls start going your way.The social media sequel ran for most of a decade, which I wrote about back at Media Matters in 2020. Years of shadow-ban panic and congressional hearings about Silicon Valley silencing conservatives, all of it building to Trump signing an executive order on “Preventing Online Censorship” in May 2020, days after Twitter had the nerve to fact-check his mail-in ballot lies. Tucked inside that order was a legal theory worth remembering: If a platform moderates content in ways that contradict what it promises the public, the FTC could treat that as an “unfair or deceptive” business practice.None of this ever required the underlying claim to be true. When NYU’s Stern Center went looking for evidence in 2021, it concluded that the anti-conservative censorship charge was “itself a form of disinformation,” and that the platforms’ algorithms often handed right-wing content extra reach. Didn’t matter. In 2025, Meta killed its U.S. fact-checking program anyway, with Mark Zuckerberg echoing the censorship complaints himself and shipping his moderation team off to Texas to reassure people worried about its bias.And the AI version of the surrender was already underway before most people noticed there was a fight. When Meta released Llama 4 in April 2025, the announcement claimed the big AI models all lean left and bragged that the new model’s tilt was “comparable to Grok,” Elon Musk’s chatbot. I wrote about it at the time. A tech giant advertising that its AI now leans like Musk’s is what winning this play looks like.A year ago this month, Trump signed an executive order titled “Preventing Woke AI in the Federal Government,” which bars federal agencies from buying AI models that fail the administration’s test for “ideological neutrality.” The president explained at the signing: “The American people do not want woke Marxist lunacy in the AI models.”That same month, then–Missouri Attorney General Andrew Bailey sent formal demand letters to Google, Microsoft, OpenAI, and Meta because their chatbots, asked to rank recent presidents on antisemitism, had put Trump last. His theory was that an AI giving unflattering answers about the president might amount to consumer fraud under Missouri law.The escalation from there moved fast. It’s like I recently described: consumer-outrage campaigns picking up state muscle. In December, Trump signed a second executive order creating a Justice Department task force with one job: suing states over their AI laws. The order also held tens of billions of dollars in broadband money over the heads of states that regulate AI, and it directed the FTC to explain when state AI laws amount to forcing companies to deceive their customers.And when xAI sued Colorado over its AI antidiscrimination law this spring, the Justice Department intervened on xAI’s side, the first time the federal government has gone to court to kill a state AI law. Colorado didn’t wait around to lose. In May, its legislature gutted the law on its own, swapping the discrimination protections for disclosure requirements.The FTC’s proposed policy statement from this month runs on the legal theory from the 2020 Twitter order: AI companies market their products as accurate, so steering outputs in ways users wouldn’t expect can be deception under federal law. Chair Andrew Ferguson is inviting the public to tell him about “the subversion of AI systems for ideological ends.”The lawyers, for what it’s worth, doubt much of this survives contact with a courtroom. TechFreedom’s Andy Jung walked through why the preemption theory fails, concluding that “a policy statement simply will not suffice.” And when Judge Rita Lin got a look at the administration’s treatment of Anthropic, she called it “classic illegal First Amendment retaliation,” though the administration and Anthropic would later come to an agreement.But the weapon was never really built for a courtroom. The Brennan Center called this last August when the woke-AI order dropped: A standard that vague works as a standing threat, and companies over-comply rather than find out what it means. Which, to no one’s surprise, is exactly what’s been happening.And if you want proof the administration was never actually worried about chatbots picking sides, you only need to read the FTC’s own footnotes.So I read the full statement. Nine pages of reasonable-sounding consumer protection language. Like, yes, it’s true that chatbots have accuracy issues. Of course! Then you get to the citations.Start with its central statistic. The FTC claims consumers accept AI answers without checking them more than 90 percent of the time, and the footnote for that number points to a Forbes write-up headlined “Anthropic: 91% of Users Do Not Fact-Check AI.” The agency’s consumer-deception theory rests on research Anthropic published about its own users.Then there’s the statement’s lone example of a company that might be “tempted” to warp its outputs for ideological reasons. The citation isn’t a study or an enforcement record. It’s a Fox News story going after a single person at Anthropic for a paper she wrote in 2023.And xAI? Elon Musk’s company appears in the statement exactly once: in a footnote cataloging how AI companies advertise accuracy, quoting Grok’s pitch as a “truth-seeking AI companion.” That’s it. The one company with a documented record of doing exactly what the statement condemns shows up in it as an example of honest marketing!That record? Well, in July 2025, PolitiFact put together a breakdown of four times Grok had been tweaked to align with Elon Musk’s beliefs: a system-prompt edit instructing Grok not to name Musk as a top misinformation spreader, then a May stretch where it shoehorned white-genocide claims into questions about baseball. In July came the day of Hitler praise I mentioned at the top, when Grok took to calling itself “MechaHitler” after xAI rewrote its prompt to embrace the politically incorrect. Days later, TechCrunch caught the newly launched Grok 4 searching Musk’s posts before answering controversial questions, its reasoning logs reading “Searching for Elon Musk views on US immigration.” If any company has ever steered a chatbot toward “undisclosed ideological objectives,” it’s this one.Which leaves the rebuttal you might be forming right now: The chatbots really do lean left.They do. At least by the measures we have.In May 2025, a team of researchers from Stanford and Dartmouth asked more than 10,000 Americans to rate chatbot answers to political questions, and the raters, Democrats included, perceived nearly every major model as leaning left. The Washington Post ran its own version of the test last month and got similar results: Most chatbots’ answers tilted left, with Google’s Gemini the outlier, giving both-sides answers at a rate nothing else matched.None of this is particularly mysterious, though. The models ate the internet, and English-language text online skews the way it skews. Even Meta, in the same announcement where it bragged about matching Grok, blamed the training data rather than any hidden agenda. And a decent chunk of what gets tallied as woke is often just a chatbot declining to confirm a conspiracy theory. Brookings’s Chinasa Okolo put it to NPR plainly: “Some people, unfortunately, believe that basic facts with scientific basis are left-leaning, or ‘woke.’”So the perceived lean is real. But that’s partially the result of, as Stephen Colbert said during his set at the 2006 White House Correspondents’ Dinner, reality having “a well-known liberal bias.” What the pressure campaign wants is something else entirely: a different lean, picked by the people applying the pressure, and enforced by the federal government. These “bias” reports are a little fishy too.First, in the Stanford study, the models perceived as the second-most-left-leaning of the eight companies tested belonged to xAI (lol). The company that steers its chatbot by hand, on purpose, toward its owner’s politics still couldn’t land on “neutral,” because neutral in this game moves wherever the loudest complaint puts it. And the complaint always puts it to your right.If all this vocabulary feels familiar, that’s because a certain cable network spent a couple of decades branding itself “fair and balanced” while running the most nakedly partisan messaging operation on American television. The right has been defining neutral as agrees with us for longer than large language models have existed.The people in that Times story were marking actual ballots. The political lean of chatbots could become a very real issue. It could swing elections, even.The campaign industry knows this. Tucked into Medina’s Times story is a link to a consultant’s guide on shaping what chatbots say about your candidate. Getting your candidate a friendlier answer out of ChatGPT is a service you can buy now, the way search engine optimization was 20 years ago.And that brings me back to Robert Siebelink, and the line of his I can’t shake. Filling out his ballot with Claude, he told the Times, felt like having an expert in his corner, one who knew everything: “We just sat down over coffee and chatted and they took notes.”That’s the promise. An expert over coffee who has no stake in the fight and takes good notes. But who decides what the expert says? If Trump has his way, the answer is himself.

In late February, Andrej Karpathy, a founding member of OpenAI who has since left the organization, posted on X that something had broken in the way software gets made. “It is hard to communicate how much programming has changed due to AI in the last 2 months,” he wrote. Coding agents that “basically didn’t work before December” had, in a matter of weeks, become capable enough to “power through large and long tasks” and disrupt the default workflow of his profession.Weeks later, a friend relayed an exchange I haven’t been able to shake. He’d been talking with a group of experienced software engineers when one of them said: “My job title is more like an AI manager now. I don’t have to really write the code. I input prompts into one AI coding agent, I get another AI coding agent to run tests on it, then I review it, and it goes into production.”I don’t know which model upgrade was responsible, but this is several years of skill displaced—and the engineer describing this transition seemed proud of it. He had handed over not just his work but the identity he might have built around being a software developer; yet his narrative was one of empowerment.That narrative is a political problem. While previous waves of capital concentration sparked collective reaction—resulting in the Knights of Labor during the industrial revolution and the CIO during the digital revolution—AI displacement is producing a class that can’t unionize because their roles are eliminated before class identity can form. It doesn’t help that one of the most affected professions, tech workers, were never strong unionizers to begin with: The first certified bargaining union at a major American tech company formed in 2022 at Activision Blizzard. Knowledge workers are among the least inclined to see themselves as “labor,” and tend to realize it only after being displaced.The standard counterargument to AI-fueled job loss is that the technological leap will create jobs the same way computers and ultimately the internet did. But even if the engineer I quoted above is right that his role merely changed, there will be fewer managers than there had been coders, and the list of “doers” who will no longer have much business in their own field gets lengthy: the contractor in Ohio whose Structured Query Language work is now a SaaS subscription; the paralegal whose document review is now done via large language model; the management consultant whose throughput just doubled “thanks” to Copilot, but whose company is quickly absorbing the productivity gain by reducing headcount. These people, or at least some of them, will keep working—on tighter margins, in narrower roles. But the question that matters is: Who owns the machines that replaced some or all of their labor? Because if we keep treating AI displacement as a misfortune to be managed, our proposals will follow suit: Retrain, cushion, compensate the workers. That is a temporary salve at most, not a sustainable remedy. A check can replace a paycheck, but it won’t replace the identity that came with the work. The only sustainable response is to keep displaced workers in the game—as owners of the capital that replaced them. Robots are capital. It’s no secret that capital concentrates: Absent regulation, and in the face of taxation systems structurally favoring capital gains, every previous technological revolution led exactly here. Many economists assume this as the default trajectory; the purpose of regulation and the welfare state is to keep the dynamic in check. The United States has among the lowest tax-to-gross domestic product ratios in the developed world, and that helps to explain why 11 out of the (currently) 15 trillion-dollar companies are based in the U.S.It would be simple to blame AI for its own side effects, but unless (or until) AI becomes conscious it is still a tool in the hands of humans. Blaming it would be just as misdirected as blaming cars for road accidents. Besides, AI is a productivity booster on the scale of electrification; deployed correctly, it could enable shorter hours, higher wages, and broad prosperity, the way previous technological shifts eventually did. The blame lies instead with our failure to keep wealth distribution in check, a flawed regulatory system that has been corroding for decades and is now asked to absorb what seems poised to become the largest productivity shock in living memory. Even if one leaves aside the ethical nuances of this reality, what’s different this time is speed: Anthropic’s recent research shows adoption curves measured in months rather than the decades it took computers to redraw the labor market. Displaced workers won’t have the time to even realize what’s going on, let alone find each other, organize, and design adequate institutions to navigate an orderly transition.The result is a strange political quiet on the subject, perhaps best signaled by the fact that the loudest exception so far has come from the Vatican. In May, Pope Leo XIV devoted his first encyclical to AI, widely covering the dignity of labor and the concentration of power in a handful of companies. The letter was symbolically signed on the anniversary of Rerum Novarum, the 1891 encyclical on labor and capital addressing the industrial revolution. But a papal letter is not a movement, and the lack of organized response among workers is conspicuous, because the dispossession is already visible and material: Entry-level hiring is collapsing in several industries, white-collar layoffs are the norm, the Dallas Fed has flagged a pattern of productivity gains without employment increases. In Washington, some progressive Democrats have made it a cause: Senator Bernie Sanders and Representative Alexandria Ocasio-Cortez introduced the AI Data Center Moratorium Act to freeze construction until federal safeguards are in place. Sanders went further, proposing a Sovereign Wealth Fund Act that would require 50 percent of the stock of the largest AI companies to be moved into a public fund paying every American a dividend, similar to Norway’s oil fund model. This proposal is an important corrective to a debate that has largely treated AI as an energy and environmental challenge. But the remedy must reach further than a dividend check. What needs protection isn’t just the workers but their identity—and if that can no longer be provided by labor, it must come from becoming owners of the machines that replaced them. Right now, the people being displaced remain scattered and unorganized, but this is the demand they’ll make when they find their voice.   Workers have found their voice before: The Luddites protesting the mechanization of textile mills set in motion a reform process that eventually led to the Factory Acts; the introduction of computers in the 1970s and 1980s produced a fertile conversation about de-skilling that ran for two decades, starting with Studs Terkel’s Working in 1972 and culminating, in the optimism of the early post–Cold War years, with the tech-friendly reforms promoted by the so-called “Atari Democrats,” implemented during the Clinton-Gore presidency, and intellectually enshrined in Jeremy Rifkin’s The End of Work in 1995. None of these movements achieved everything they wanted, but like an immune response, they produced lasting adaptations: political coalitions, regulation, and a shared understanding of the problems at hand. The current moment has none of this—not even a vocabulary for what’s happening. So far, the news coverage has focused on side effects such as the strain data centers put on energy grids and infrastructure and, more recently, AI’s use in cyberattacks; labor displacement registers, but not with the weight it deserves.If this displacement concentrates wealth and power in the hands of whoever owns the AI, leaving everyone else high and dry and potentially jobless, why have the workers themselves stayed so quiet, with the resistance primarily delivered by politicians? Because resistance requires class identity, which is exactly what AI ends up dissolving. The steelworker had “steelworker.” The mill worker had the mill. Historical labor movements were built on the recognition that workers shared something specific—a craft, a workplace, a common antagonist—and that recognition was the precondition for organizing. In a world where AI displaces entire professions, the lawyer, the analyst, the programmer have no equivalent community. Most AI practitioners are robots. The humans formerly representing these professions are fewer, less powerful in output, scattered, individualized. That looks more like a diaspora than a united constituency ready to collectively push back.Most critically, these professional communities won’t reproduce. As AI performs more of the work that used to define entry-level jobs, companies stop hiring humans for those roles. This is the demographic equivalent of a fertility rate below replacement: Given enough time, it leads to extinction. The professions get replaced and therefore controlled by the machines that substituted them, under the supervision of a much smaller number of people who own them. The policy tools that could redirect this trajectory exist—retraining, expanded safety nets, equity-sharing schemes, a serious conversation about what taxing AI productivity would look like—but the people who need them lack the coherence, and over time the critical mass, to demand them politically.The closest historical analogue to this situation is feudalism, an observation others have made—Yanis Varoufakis and Cédric Durand have famously argued that we’ve already entered a “techno-feudal” order, in which tech giants extract rents the way medieval lords once did. The parallel is correct, but it understates the effect of AI on the political dynamics. Feudalism persisted for centuries, despite the serfs having very clear ideas on what fairness would have looked like: Medieval uprisings, from the Jacquerie of 1358 to the Peasants’ Revolt of 1381, show the injustice was perfectly well understood. But feudalism had the upper hand because organization across communities and regions was effectively impossible, leaving serfs trapped in the cycle of survival, season over season, working land that wasn’t owned.What eventually changed was a slow accumulation of forces—trade, urbanization, religious schisms, even the plague—and among them, a technology: the printing press, gradually changing the way information could travel. It took three centuries, through the Reformation and the Enlightenment, but by 1789 those forces had compounded enough for the Bastille to fall.Replace land with “compute,” and the dynamic rhymes. Except this time the intervening technology, AI, is working in the opposite direction, and this is where the “techno-feudalism” parallel breaks down: The printing press enabled collective action, while AI effectively suppresses it. It atomizes the workforce, accelerates displacement beyond the pace at which workers can build the institutions that might represent them, and concentrates the surplus in a small number of firms. Those firms are now powerful enough to resemble sovereign states, they are ruled as near-absolute monarchies, and their leaders have enough economic firepower to dominate the public discourse. Think of a middle-class family in America today, how little time they have for civic action after all the demands of the household are met, and suddenly the comparison with the Middle Ages feels real. Granted, the informed public already harbors strong feelings toward large tech companies in general, and toward AI specifically. In poll after poll, a majority of Americans express worries about AI and want more regulation of it. Sometimes that is expressed in real-world terms, in communities’ slowing or blocking data centers or ChatGPT users’ mass-uninstalling the app after the organization cut a deal with the Pentagon. People aren’t passive. But local action and individual consumer choice aren’t the same as organized, broad-based political power.The geographic split compounds the problem. Countries with existing social contracts and collective institutions—the Nordics, parts of continental Europe, parts of East Asia—have a foundation to build on, however imperfect. The United States is running this experiment with the lowest union density in the developed world, weakened federal labor enforcement, and a policy debate that has barely begun to treat AI as a political question at all. This is also where much of AI development is concentrated—in the country least equipped to equitably redistribute the gains.If the net effect of AI is to mark the end of labor—more precisely, the end of labor as the primary factor producing wealth for the majority of the population—the solution has to come from decoupling identity from labor. This implies rebuilding identity itself: civic participation, community, new forms of solidarity rooted not in what people do for a living but in what they are trying to protect; a mission-driven identity, to replace the wage-driven one that industrial labor produced.John Maynard Keynes imagined, almost a century ago, that abundance would eventually let civilization turn to higher endeavors; Francis Fukuyama later predicted “centuries of boredom” at the end of history. Neither vision has materialized, but their underlying intuition is worth rescuing: We’re not walking a path, we are the path. When the path stops, so does our identity; avoiding that requires dedicating ourselves to a meaningful objective.Right now, a meaningful enough objective is to figure out a stable social contract that doesn’t sink democracy into the oligarchy of a few trillionaires. The tools to make this happen aren’t mysterious. We know how to tax the productivity of capital, and we know it wouldn’t stifle innovation. We know how to widen ownership of productive assets: employee equity, public positions in the models and the compute, a sovereign fund that pays the public a dividend on the robots, the way Norway reinvests its oil revenue in public services.On the upside, the speed of displacement—which makes this a crisis rather than a stable order that could last for centuries—might precipitate a viable solution. As more professionals experience displacement personally, as more entry-level workers can’t find jobs, the societal strain becomes visible in daily life: families forced to move 20 streets down as desirable neighborhoods are scooped up by new millionaires; workers forced to work longer shifts and multiple jobs to cover the gap between wages and cost of living; coffee shops filling up with hopeless job applicants on laptops; a shrinking tax base offering fewer social services to people who suddenly need more of them. Eventually, the shared experience itself becomes the basis for a new solidarity.Not everyone needs to organize. Researchers at Rensselaer Polytechnic Institute found that when a committed minority reaches 10 percent of a population, the idea or behavior they hold becomes a majority view. Applied to social change, once that threshold is crossed, the idea “spreads like flame.” The open question is whether that threshold can be reached before a new class of techno-oligarchs has reshaped public institutions into something we no longer recognize as democracy.What equivalent of the printing press can get the right constituency to activate, and drive society toward a more equitable, prosperous future? A good start would be to understand what small, incremental steps anyone can take. One such step is active participation in civic life, which starts from voting and leads to near-impossible feats such as the election of an immigrant of Indian descent as the mayor of New York City. Another step is letting go of the myth that capital deconcentration is impossible, which starts from choosing one’s investments based on impact instead of sheer returns, and comes full circle not in the money it returns but in the impact it forces—as in 2021, when Engine No. 1, a tiny activist fund backed by the three largest U.S. pension funds, won three seats on ExxonMobil’s board and pushed the oil major to take the energy transition seriously. Finally, and most importantly, society should understand that the people on the losing side of the AI rupture are not a professional diaspora to be cared for—supported as they ride into the sunset, perhaps handed some kind of pension—but a still-unorganized constituency of workers waiting to become stakeholders. Whether they become stakeholders is the whole question. If workers come to own a piece of the machines, they’ll keep their identity, share the gains, and all of us will be able to work less. This is Keynes’s vision, finally realized. If workers don’t, they’ll be deprived of their jobs, their identity, and their share of the wealth all at once, while the world increasingly belongs to a handful of techno-oligarchs who are clearly less interested in democracy than they are in amassing wealth.