Artificial intelligence is rapidly reshaping American society, from the workplace and the military to schools, financial markets and local communities. While tech companies promote AI as a revolutionary force, growing concerns about its risks are colliding with a massive expansion of data centers, rising electricity demands and fears of economic disruption.
In their Left & the Law segment, Jan Haaken and Mike Snedeker examine the limits of AI regulation, the legal accountability of tech companies, the human costs of AI-driven military decisions, and the financial mania fueling the industry. They also explore the growing community resistance to data center expansion and whether a moratorium could create space for a more democratic debate about who benefits from these technologies—and what kind of society they should serve.
AI Apocalypse: Real Danger or Corporate Hype?
Jan:
The regulation of AI has moved from the periphery to the center of American politics in recent months, as AI executives themselves warn that their technology could advance beyond human control and threaten to wipe out all of humanity. While some critics believe this is mere hype promoted by AI companies, others suggest that the risks of advanced AI are already evident in documented cases, such as AI agents hacking into companies and military attacks based on AI-selected targets that turned out to be mistaken.
There are also more everyday risks, including the undermining of young people’s intellectual and social development.
Mike, how would you approach legal strategies for confronting the problems posed by this technology, which has invaded every corner of our digital lives?
Mike:
The world has indeed been transformed, and will continue to be transformed, by AI. But it’s been transformed before. Society has undergone major transformations repeatedly, and methods already exist for dealing with AI excesses and abuses. It boils down to recognizing that AI is a human creation.
The apocalyptic predictions are just hokum. There’s a way of claiming authority by saying, in essence, “This is too complicated. Only brilliant people like me are capable of understanding it, so leave all its development to me.” That’s nonsense.
Regulating AI Without Surrendering to Big Tech
Jan:
When confronted with that nonsense, some people say, “You’re right. There needs to be federal oversight or federal regulation of these technologies.” But others argue that this is precisely what these companies want from the government: a kind of firewall for themselves, allowing them to pretend that something meaningful is in place on the regulatory front.
Mike:
I think that’s right. The creation of a new regulatory agency run by someone like Sean Duffy or Robert Kennedy Jr. would do nothing but impede efforts to hold AI companies accountable for their mistakes. There are already eight or nine federal agencies that govern different aspects of this technology. Actual experience shows both their limitations and their strengths.
Jan:
Before we get into those federal regulatory agencies, I think part of what you’re suggesting is that the idea of creating a sweeping, unified AI regulatory framework accepts some of the industry’s own premises: that these technologies have godlike powers, have escaped human control, and have a life of their own.
That’s part of the broader mythologizing of superintelligence: the idea that these technologies can become entities that operate independently of human agency. The industry argues that we need a superagency to control this supposed superintelligence.
Mike:
We’ve been around this track before. Consider railroads, electricity, and the internet. These have all brought enormous transformations to society, with significant casualties along the way. Bad things have happened. But, slowly and steadily, civil and criminal systems for dealing with these problems have developed around the world.
AI companies say they know full well what they’re doing, even as they make these apocalyptic predictions. Those claims could come back to haunt them when lawsuits arise from AI abuses that lead to a traffic or banking fiasco, or something else disastrous. Something bad will happen. I’m surprised that nothing truly terrible, beyond what happened in Iran, has resulted from AI’s pervasive presence.
But we have civil laws. We have criminal laws. We can arrest people. We can sue them. And in some parts of the AI world—which now encompasses the entire world—companies have already been sued.
Jan:
Before we get to the regulatory agencies currently operating within the federal government, can you say a bit more about the lawsuits that have been brought against these companies?
When AI Fails: Corporate Negligence and the Human Cost of Automation
Mike:
The Federal Trade Commission, or FTC, essentially oversees consumer protection. For a while, it was led by Lisa Khan, one of the best officials in the Biden administration.
In December 2023, Rite Aid was banned from using facial recognition technology for five years after an investigation found that its AI surveillance system had falsely identified customers as shoplifters, disproportionately targeting women and people of color. The FTC found that Rite Aid had deployed the technology without reasonable safeguards and had failed to prevent harm to people who were wrongly accused. Dozens of people were falsely identified. They were confronted by employees and even detained based on false matches. That was an example of effective regulation.
Jan:
What was the outcome of that investigation?
Mike:
The outcome was that Rite Aid was banned from using facial recognition technology altogether. There have also been individual lawsuits of various kinds, with a range of outcomes. But the problem was clearly identified and formally recognized by a federal regulatory agency. The FTC also did a pretty good job in cases involving Weight Watchers and other consumer-facing companies. Of course, Lisa Khan was removed fairly quickly after President Trump took office.
Another example is the Food and Drug Administration, or FDA, which has regulated medical devices for many years, including those that use AI for diagnosis. Can AI review your X-rays? The FDA has been regulating such applications for some time.
But the biggest user of AI right now is the Department of Defense.
In 2017, Google launched Project Maven, a project that used AI to analyze drone footage. Many Google employees quit because they objected to the use of machine learning for military purposes. The technology is now almost a decade old, and Google eventually withdrew from the project. Palantir, however, did not.
Palantir has been paid handsomely by the Trump administration. Apparently, Palantir and Project Maven were involved in the process of identifying and selecting the school that was bombed on the first day of the war against Iran, killing 180 schoolchildren. How did that happen?
The military said it had relied on AI to identify the school. But Admiral Brad Cooper, head of U.S. Central Command, said humans would always make the final decisions about what to shoot, what not to shoot, and when to shoot. He also boasted that advanced AI tools could turn processes that once took hours or even days into seconds. He’s proud of how effective their AI is, yet seems to overlook the consequences of moving so quickly, even after all the bloodshed.
I think the problem there wasn’t the AI itself. The problem was human beings who lust for an immediate resolution, just as they did in Venezuela, and who want to bomb, bomb, bomb and win right now. That’s a human characteristic. It doesn’t really have anything to do with AI.
AI and the Future of Work: Who Wins and Who Gets Left Behind?
Jan:
But it does have to do with both the military logic of empire and capitalist conceptions of efficiency, including the elimination of workers who possess valuable skills. The entire history of the Industrial Revolution involved eliminating workers who held the knowledge, skills, and craftsmanship of early industry and breaking production down into interchangeable parts so that people could be replaced. That was part of Taylorism and Henry Ford’s model of production.
These technologies are now being used in the name of efficiency, but efficiency is also viewed through the lens of profit and capitalism’s drive to generate a surplus. There are powerful economic imperatives driving this industry and shaping how these technologies are used everywhere.
Mike:
Productivity did continue to increase during that period because someone had to design and operate the new systems. The number of new jobs created kept pace with, or even surpassed, the number eliminated. I think that may well happen here, too.
We’re talking not only about the assembly-line jobs discussed by Frederick Taylor in the development of Taylorism, but also legal assistance, financial advising, and even white-collar jobs that are now endangered. A host of new jobs will be created, I suspect in comparable numbers, but the transition will be rocky. The shift from the past to whatever the future brings will not be smooth.
Jan:
That was part of Clinton’s message to workers in the Midwest in the late 1990s. He said, in effect, “Look, I can’t bring factories back to these dying industrial towns. They’re not coming back.” He gave speeches along those lines. But he said he could provide education to help people train for the new technological society and economy.
Much of the anger today in the Midwest and in more conservative parts of the country stems from the false promise that there will always be a new technological savior. Many people know they’re going to be left behind.
The data centers that support these technologies have consumed enormous amounts of land and resources in parts of the country without delivering on the promise of new jobs. You’re suggesting that the technology could be used differently, in a way that’s more consistent with a progressive vision of a good society.
Mike:
Clinton wasn’t making a technological argument. He was covering up for the fact that he was a spokesperson for big capital, which regarded labor as a cost—as garbage, as something to be sourced wherever cheaper workers could be found around the world. This wasn’t a technological issue at all.
He could have built factories using advanced techniques right there in the Midwest. But treating labor as an inconvenient cost is what eviscerated America’s industrial heartland and stripped manufacturing jobs of their prestige. Today, nobody wants their son or daughter to grow up and work in a factory. Yet 60 years ago, that was a perfectly respectable and prosperous job in the United States. The same possibilities exist today. The same factories could be built, but the values of money are prevailing.
Coming back to the regulatory agencies, your point about data centers is important. I think it’s a big part of why AI is a major issue in China without being quite as overwhelming there as it is in the United States. China has cheap electricity. It has steadily developed solar and wind power without eliminating coal-fired electricity. As a result, electricity is relatively cheap and abundant.
That’s not the case here. For decades, there was little incentive to develop major new sources of electricity. Now we’re being hit simultaneously by AI and the construction of enormous data centers, all of which require vast amounts of electricity. And they require one hell of a lot.
Jan:
I think data centers also embody some of the same irrationalities of capitalism that I was talking about. These enormous, windowless structures seem to be colonizing the landscape everywhere. They’re extremely inhospitable. There are very few cars outside, very few workers, and yet they extract enormous resources. They also rely on secretive deals and nondisclosure agreements with local politicians.
This fits into a broader political analysis that more people are beginning to recognize: the people running things aren’t just the corrupt Trump administration, his family, and his allies in big industry. Much of the system we live under is driven by irrational economic and political forces that work against the well-being of communities.
The Data Center Boom and the Hidden Costs of the AI Race
Mike:
It’s become a mania, Jan. I think that’s the source of the problem. It’s a financial mania.
An enormous amount of money is involved. In fact, a huge portion of the American economy now hinges on building these data centers. And, by the way, they’re not lofty things floating somewhere in the cloud. They involve thousands of tons of steel and copper. As you said, they’re windowless. They brood there. They consume monstrous amounts of electricity.
Because there’s a slight increase in efficiency with every 10,000 new chips you buy, there’s a maniacal increase in funding. Even though the largest companies are generating more and more revenue—for example, Anthropic brought in $4 billion last year, but its expenses were $8 billion—there’s a huge, crazy surge of money, time, interest, and attention devoted to building these hyperscale AI centers. At the same time, AI is increasingly pestering people in their daily lives.
Jan:
The whole thing has a certain mania to it. It’s like musical chairs: the music keeps playing, everyone keeps moving, and at some point, everyone’s going to have to find a seat. I don’t know if that’s the right analogy, but eventually the music will stop.
As you said, capitalism is characterized by bubbles and cycles of boom and bust, after which things shake out again. But this bubble carries risks that apparently exceed those of earlier economic bubbles.
In the meantime, people are organizing in their communities. I think you’re suggesting that organizing around data centers should happen at the local level, including calls for a federal moratorium until we better understand what’s happening and have more democratic input into how these projects develop. People are confronting something central to the economic system we live under.
Mike:
Yes. The pace is breathtaking, and it’s problematic. Data centers are being built, but so are the electrical lines needed to supply them. Years of delays and backlogs in electricity development have collided with the rapid expansion of data centers, all of which is happening around us.
Those building these facilities are taking shortcuts wherever they can because they know they’ll face resistance. There’s deception in the air, along with a maniacal pace. Transparency is declining in the construction of both data centers and new electrical lines. So I think the idea of stopping and taking a deep breath is a very good one.
But what would happen to the American economy? People say the wheels of the merry-go-round have to keep turning. If we stop, the world will end. I don’t think the world will end, but there would be disruption. At this point, though, there’s no real alternative. We need to put things on hold for a while and digest everything that’s happened before moving forward.
Jan:
Okay, Mike. I think we’ll see more of this story unfold. These tech industries have had plenty of airtime to tell their stories, both the utopian and dystopian versions of their products. But there are other important stories emerging about how people are coming to terms with these developments and finding solidarity in their communities. That includes solidarity with Native American people, who have been fighting these kinds of colonial frontier projects for centuries.
We’re seeing solidarity in places where it hasn’t always existed before, including among labor groups, environmental groups, and Indigenous communities. I think there’s hope, even in hard times.
Mike:
These kinds of efforts are essential because the tech industry is swallowing up not only everyone’s time but also everyone’s money. A huge portion of the stock market’s substantial gains over the past few years—roughly an 80% increase—has been driven by just six or eight companies. Their frenzied, circular financing is repeatedly driving up their stock valuations. They’ve absorbed enormous amounts of attention and resources, affecting virtually every aspect of modern American life.
The more we think about it and talk about it, the more convinced I am that we need some kind of moratorium. We need to stop, take stock of where we’ve come, and ask ourselves where we really want to go.
Jan:
Okay, well, thanks, Mike. We’ve got some elections coming up in a few weeks, so we’ll see whether progressive Democrats can take this on if they regain control of Congress. Addressing the military and the irrationalities that drive military uses of this technology will certainly be part of that.
Mike:
One extraordinary fact about this election is that the only issues on which there’s bipartisan consensus are calls to stop these data centers.
Jan:
Yeah, that says something. The industry has been pushing this grand narrative about super-entities and superintelligence. There’s often an unintended consequence when you overhype something.
This is Jan Haaken and Mike Snedeker, speaking today on The Left and the Law about legal strategies for resisting AI.
💬 Share Your Thoughts
Should governments impose a moratorium on new AI data centers until their environmental, economic, and social impacts are better understood?