AI is already eroding privacy, undermining public services, and imposing costs on local communities. Experts say strong regulation could help prevent further damage and ensure the technology’s benefits are shared more broadly.
This article was made possible because of the generous support of DAME members. We urgently need your help to keep publishing. Please join us for as little as $1.00 a month or make a one-time gift in any amount!
Kevin De Liban, a civil rights attorney, championed people with disabilities in Arkansas. Thanks to a Medicaid program, De Liban’s clients were regularly visited by state-funded caregivers who helped them bathe, get out of bed, go shopping, and cook meals.
But that care slowed to a trickle in 2016. Some of De Liban’s clients — people with cerebral palsy and quadriplegia — were suddenly without care. De Liban said clients dependent on scheduled caregiving for basic needs were literally left alone, “lying in their own waste and suffering from bed sores.” The State of Arkansas, De Liban soon learned, had implemented an automated computer algorithm to decide how much home-based care to provide people.
De Liban sued, ultimately convincing the Arkansas legislature that letting a machine decide who gets what care was “too cruel, even for them.” After the suit, De Liban said he started hearing from people across the globe. He realized that “this technology is being deployed against low-income people in all kinds of really harmful ways.”
Ten years later, and it seems almost every aspect of our lives has been subsumed by AI. In little ways, like a simple Google search, and in big ways, like war. The U.S. military, for example, used AI to wrongly identify an Iranian elementary school in 2026, which it bombed twice, killing over 100 little girls and over 25 school teachers. Elsewhere, AI data centers are being unconsensually forced into communities, sucking up already strained water supplies and even displacing American families from their homes. In places like Northern Virginia, Texas, Tennessee, and Utah — where some of the world’s biggest data centers are located — residents deal with relentless, 24/7, high-pitched, industrial whining sounds. They also must contend with higher utility bills and less water (data centers eat up 20% of Virginia’s total electricity, for example; they also, on average, consume over 100 million gallons of water each year).
Data centers make for unsightly neighbors as well. They are dull-colored, hulking, windowless sprawls of cooling towers, gas generators, and utility lines. A New Jersey resident who lives next to one such data center told a reporter: “I don’t want to move,” but, “who is going to pay what the house is worth with that thing right there?”
Aside from being an eyesore and a serious ecological problem, AI is helping to turn George Orwell’s nightmare of an authoritarian surveillance state into a reality. The technology has augmented ICE’s ability to surveil, hunt down, and then abduct immigrants. It powers the Flock Safety cameras that watch us from above. The Feds have used it to infiltrate activist circles in the U.S. and coordinate drone strikes in Iran.
Reality itself has been warped, too; AI-generated content has flooded social media, making it harder to discern what is real and what is not. It is automating the workforce, sterilizing human interactions, and plucking the essential element of art — humanity — out of literature, poetry, animation, photography, graphic design, voice acting, and filmmaking.
Everyone, it seems, is using it. In 2024, Pew Research found that about 33% of Americans routinely used AI chatbots. Two years later, in 2026, that number had climbed to about 50%. I met a college student the other day who told me that they use AI to craft “better” text messages to send to their friends and family. Even the president of the United States is an avid-user (recall the time Trump posted an AI-generated video of him flying a fighter jet over the “No Kings” protests, onto whom he later dumped feces).
All this seems, well, horrible. Or, as De Liban would describe it: “one-sidedly awful stuff.” But what about AI being used to improve the human condition? What if, for instance, AI could help us cure disease? De Liban is skeptical. He says that regardless of AI’s disease-curing capabilities, “the problem is the incentive structures aren’t in place to ever create good uses of AI at scale. There’s no political accountability when these things go wrong, there’s very little market accountability, and there’s very little legal accountability. And so when you have this vacuum of accountability, all the incentives are to do wrong, to do harm.” Even if AI was used to somehow cure a major illness, De Liban said, “there’s the question of who would be able to access the treatment that AI generated if AI is also cutting your access to health insurance.”
Among the experts I spoke to, there is overwhelming consensus that federal regulation is what is direly needed to reign in these Big Tech companies that’ve unleashed AI. Though that isn’t happening. In September 2026, Trump and executives from OpenAI, Meta, and Google signed a deal that will allow tech companies — not the government — to police their own AI development. Meanwhile, the Trump administration is pouring billions into AI research and expediting federal approvals for data center projects all over the nation.
The ones sounding the alarm bells aren’t merely a small handful of zealous activists. Interestingly (and perhaps disturbingly), tech executives themselves are issuing grave warnings about the technological Pandora’s box that they opened. The guy responsible for ChatGPT, OpenAI CEO Sam Altman, has said things that make it seem as though he is worried about AI. Per his own words, Altman appears concerned AI will invent new diseases, spread false information to millions of people, bolster authoritarians worldwide, crash economies, and make the rest of us dumber as it gets smarter.
“What I lose the most sleep over is the hypothetical idea that we already have done something really bad by launching ChatGPT,” Altman said at a recent Economic Times event. “That maybe there was something hard and complicated in there that we didn’t understand and have now already kicked it off.”
Who Holds AI Accountable?
Thomas Hofweber, a professor of both philosophy and data science at the University of North Carolina at Chapel Hill, prophesies two very distinct futures for the U.S., each hinging upon how we collectively decide to address AI.
The dark future, as he explained it to DAME, is one where wealth is even more concentrated and unevenly distributed than it is now. AI will have ushered in an oppressive surveillance state, and the very few who control the majority of the wealth and AI “will run the country.” That’s if we don’t act, Hofweber thinks.
If we do something about it, through strict regulation and international cooperation, Hofweber can envision a future where AI is developed safely, and the benefits are reaped by the many, not just the rich few. Scientists, aided by AI, could make breakthroughs in research, possibly even developing cures for previously incurable diseases — and those cures are made available to all. Negative uses of AI, like mass surveillance, are outlawed. “Quality of life goes through the roof,” he imagines.
If we are to avoid the first option and achieve the second, Hofweber says, “It is not so much an issue about areas of deployment than an issue of how it is deployed in particular areas.” As an analogy, he said, “having technology as an intermediary in close interpersonal relationships might seem like a bad idea, but having a FaceTime call with your partner is a good idea. But replacing your partner with AI is probably a bad idea. It just depends on the particular situation and implementation, not so much the area.”
To get to something resembling Hofweber’s second vision for the future, we need regulation. It’s unnerving, then, that we are unlikely to see any substantive calls for regulation coming from the White House. That doesn’t mean there aren’t some promising things happening in government, however.
A bipartisan bill was recently introduced in Congress to establish federal standards to monitor and verify autonomous AI agents. The bill would also give the government a “kill switch” so it could quickly shut down AI agents that go rogue. More aggressive bills, like Sen. Bernie Sanders’ proposed Ban Artificial Superintelligence Act, calls for an immediate freeze on all development of advanced AI until the appropriate federal regulations can be enacted.
What we can do is vote for political candidates advocating for brazen action on this issue. It was a shame that Francesca Hong, the only candidate at the 2026 Wisconsin Democratic gubernatorial primary who supported a one-year moratorium on data centers, didn’t win. It’s important we take seriously candidates, like Hong, who want to actively fight this problem, and that we meet candidates who don’t with skepticism.
The Lie of AI Inevitability
In our everyday lives, there are two aspects of AI that I find especially troubling. The first is the way in which it is suddenly cropping up, without permission, in mundane facets of our daily existence, from writing emails to scrolling social media. And the second is the fact that many of us are willingly and unthinkingly accepting these changes. It reminds me of a quote from the AI-critical, satirical, dystopian film Good Luck, Have Fun, Don’t Die, which I recently watched on a plane: “Progress is only progress if it makes things better. Otherwise, it’s a mistake.”
If we are to incorporate AI into our own lives, we should carefully think about how we do this. Shall we export our writing capabilities and, by extension, our capacity to think to a chatbot? The answer, I hope, is no.
Maybe critics will call me a Luddite, but AI criticism is far more than a desperate clinging to the past. Increasingly, we are offloading our cognitive abilities by automating our lives. People use chatbots to plan vacations, decide what to eat and where, generate workout routines and diets. Some use these bots as a form of cheap therapy. Others – and this is less common – delude themselves into thinking they’ve fallen in love with chatbots. Travel, food, exercise, conversation, and love can be messy and unpredictable, but these are also the very experiences that color our lives. Delegating all this to a bot gives me mental nausea. Does this make me a technophobe? A recent MIT study, which revealed the corrosive effect ChatGPT has on our brains, might answer that question.
Kevin De Liban doesn’t believe I’m alone in my thinking. When people are exposed to what AI is really doing, De Liban says it is “wildly unpopular” across all ages, races, and even political parties. For example, social media is awash with “How to cut down Flock camera” videos. Town halls across the country have been packed with fuming residents, angry that data centers are being thrust into their communities. Despite finding that it is commonly used, Pew Research also found that “52% of Americans are more concerned than excited” about AI.
“The challenge is making sure people understand it’s doing all this right now,” he said. Once they understand all the ways AI is shaping their lives – cutting healthcare, calculating lower wages, powering surveillance — there is “immense support to regulate it.”
To regulate AI means to give communities, not solely private corporations, oversight and control in how it is implemented in society. An example is a bill introduced in 2026 by Sanders and Rep. Alexandria Ocasio-Cortez — the Artificial Intelligence Data Center Moratorium Act — that would give local residents the power to ultimately approve or reject the building of AI data centers in their backyards. As of writing, the bill remains stuck in committee. Or, in other words, legislative purgatory.
Regulation also means giving both private companies and the federal government very high bars to clear before using AI, to ensure it is safe to do so. It also means creating easy off-switches, so if AI does go rogue — as is happening on an increasingly frequent basis — people can shut it down. Regulation includes establishing tangible legal penalties for wrongdoers who use it in harmful ways. And, in some cases, it means banning it outright, as New York City Mayor Zohran Mamdani and Schools Chancellor Kamar Samuels recently did in NYC’s public schools for students below the 8th grade, and the University of Chicago Law School did for first-year law students.
Of course, it won’t be easy. “We’re up against Big Tech, and a lot of political allies from both parties, who are fine with this technology being unregulated and causing widespread harm,” De Liban told me. It’s worth noting that at least 106 members of Congress have financial investments in AI. “So it’s going to be a real fight. But through people power, that’s really where the hope comes from to overcome the money and institutional power that Big Tech has.”
Kevin De Liban is an example of someone who used his skillset — in his case, the law — to do something. After successfully battling the State of Arkansas over its automated computer algorithm cutting Medicaid benefits, De Liban went on to found TechTonic Justice, a nonprofit that conducts research, litigates, and educates regarding the dangers of AI.
De Liban said that his goal is to connect on all those fronts to “mount a really effective nationwide resistance that centers and ultimately is led by the communities who are most hurt by these technologies,” he said, referring to low-income, working-class people. “And through the process of doing that, hopefully reinvigorating American democracy and seizing it back from the forces of authoritarianism.”