← The Preamble
The Preamble
How ICE Is Mimicking 19th Century Slave Patrols, and What AI Tech Billionaires Really Want
00:45:26
How are ICE’s modern tactics echoing one of the darkest chapters in American history? Plus, AI tech executives are motivated by something bigger than money.
Loading summary
Transcript50 lines
- [00:00]
A
Foreign.
- [00:04]
B
Welcome to the Preamble Podcast. This week we're joined by Khalil Green for our lead story. Khalil is known as the Gen Z Historian and he is a regular contributor to the Preamble. He's won a Peabody Award for his work and was recently listed on the Forbes 30 under 30 social media list. This week he is looking into the slave patrols of the 19th century and how ice mimics the tactics they used. And ahead, I'll speak with Karen Howe about her book Empire of AI. She spent years investigating and found the executives behind the AI boom have motives beyond money, even the ones who claim to be doing it for good. I learned so much from this conversation. I'm Sharon McMahon and this is the Preamble Podcast. Here's Khalil.
- [00:57]
A
What's up everyone? My name is Khalil Green, AKA the Gen Z Historian, and in my story for the Preamble I'm exploring how ICE tactics today look very similar to the slave catchers of the 19th century. On June 2, 1854, the city of Boston was under military occupation. Federal troops with a loaded cannon stood ready as 1500 militiamen formed a human corridor from the courthouse to the harbor to march a single prisoner through the streets. His name was Anthony Burns and He was a 20 year old clothing store worker. His crime was escaping slavery. Shortly after Burns was first captured and held pending a hearing to confirm he was a fugitive slave, a mob gathered at the courthouse and stormed the door in an attempt to rescue him. A brawl with law enforcement ensued. By the time more police appeared and ended the riot, a federal deputy was dead and Burns remained captive. Days later, as Burns was marched toward the harbor where he was to be loaded onto a ship and returned to slavery in Virginia, 50,000 people crowded the route in protest of his removal. Storefronts were draped in black American flags hung upside down. Someone had suspended a coffin across State street with one wordpain on its side. Liberty. As soldiers beat back the crowd with bayonets, protesters shouted back at them. Shame. Authorities managed to get Burns onto the ship, but the spectacle cost the federal government $40,000 and and unprecedented manpower. It also backfired completely. Within months, Boston became a no go zone for slave catchers. According to historian H. Robert Baker, secret resistance networks formed and Massachusetts passed laws that made federal enforcement of the Fugitive Slave act nearly impossible. No freedom seeker was ever captured in the state again. Over the last few weeks, as agents from U.S. immigration and Customs Enforcement, ICE Age, and Border Patrol have flooded the streets of Minneapolis, snatching away residents and sending them to faraway detention camps and even killing citizens. Commentators have looked at history for comparisons, often reaching for World War II analogies that liken ice to the secret police of Nazi Germany. But we needn't look outside the US for the clearest parallel. ICE isn't the Gestapo. ICE is the slave catcher. And the Fugitive slave Act of 1850 tells us everything we need to know about where this is heading. Before the 14th Amendment was ratified in 1868, Black people in the United States were not considered citizens. As a black person, if you were enslaved and left the domain of your enslaver, you were breaking the law. Unless you carried papers declaring your freedom or authorization for travel, you were effectively undocumented. And any undocumented black person was subject to arrest and punishment. However, in the mid-1800s, the the nationwide political balance began to shift in favor of northern states. With California and New Mexico poised to enter the Union as non slave states, the South's grip on the Senate looked increasingly fragile. Fearing the introduction of laws that would favor the enslaved over the enslaver, Southern politicians warned that their states would leave the Union if their property rights weren't protected. South Carolina's John C. Calhoun openly threatened secession. The resulting compromise, the Fugitive slave Act of 1850, created one of the most aggressive federal enforcement regimes in American history. The new law didn't simply allow slave owners to retrieve enslaved people who had escaped. The law empowered federal marshals, officers of the United States government, not local police, to enforce it directly in northern states where slavery had been abolished. This was unprecedented. It was the first time in American history that the national government established a local law enforcement presence. The law also required state and local officials to actively cooperate with capturing fugitives regardless of whether their states allowed slavery. Beyond that, it deputized ordinary citizens to participate in the hunting by granting marshals the authority to conscript passersby to act as their impromptu assistance. If someone refused to help a marshal catch a fugitive slave, they could be fined up to $1,000. Those who interfered with the capture risked half a year in prison. Meanwhile, the legal process applied to fugitive slaves was designed for conviction. Once captured, an accused fugitive would be brought before a federal commissioner, a government appointed official, not a judge or jury, for a summary hearing. The enslaver needed only to present satisfactory proof of ownership, which could be as little as a sworn affidavit from a Southern court. Such affidavits were treated as conclusive evidence. The accused, however, was barred from testifying in their own defense, and the system was financially rigged. Commissioners earned $10 for every person they returned to slavery, but only $5 if they ruled the evidence insufficient. The law didn't just permit injustice, it paid for it. Racial profiling was the engine of slave catching in the 1800s, much as it is the engine of immigration enforcement today. Consider the story of Solomon Northup, a free black man born in New York, Northup was a farmer, laborer and talented fiddler. His father had been enslaved, but was freed after his master's death and eventually acquired enough property to vote. Solomon received an education, married, and built a life in Saratoga Springs. That all changed when two strangers approached him in spring 1841 with what seemed like a promising opportunity. Paying work as a musician in a traveling show headed towards the nation's capital. In Washington, the con turned violent. Northup was poisoned, and when he came to, he found shackles on his wrists. Traffickers moved him through Richmond and by sea to New Orleans, where he was auctioned off into slavery under the name Platt Hamilton. For the next 12 years, NorthUP passed through several owners in Louisiana's Red river region. Northup wasn't targeted by accident. He was targeted because he was black, and in antebellum America, any black person could be turned into property if no one was around to stop it. Northup had papers. He was a citizen. But it didn't matter then. And as we are seeing with the normalization of racial profiling and widespread arrests of citizens by immigration officers, it does not matter now. Historian Richard Bell even documents that a secret network of human traffickers, which he calls the reverse Underground Railroad, kidnapped tens of thousands of free black people from northern cities and sold them into slavery in the Deep South. Any black person could be seized on the street, and if they could not immediately produce documentation, or if that documentation was ignored or destroyed, they were fair game. Northup's case became famous only because he managed to get word out and secure his release. Most never did. According to a UCLA analysis, Latinos accounted for nine out of 10 ICE arrests during the first six months of 2025. An arrest nearly doubled during Trump's first 100 days in office. Community based enforcement, meaning raids targeting people in their neighborhoods rather than at the border or in jails, surged by 255% in September 2025. The Supreme Court gave this surge legal cover. Justice Brett Kavanaugh, in a solo concurrence in Noam vs Vasquez Perdomo, laid out a framework that would permit agents to weigh a person's apparent ethnicity, the language they speak, and where they're encountered as part of the calculus for a brief stop. Advocates have dubbed the resulting encounters Kavanaugh stops, brief detentions in which agents stop people based on how they look or sound, check their papers and release them if they're determined to be lawfully present in the US As University of Minnesota law professor Emmanuel Molleon called the Star Tribune, the opinion represents a major turn in the doctrine because it would essentially be adopting the position that racial profiling, ethnic profiling is reasonable under the Constitution, which no other court has said thus far. In addition, there's no official definition of what constitutes a brief stop, and there have been reports of people being held for hours or even days before being released. The Trump administration claims immigration enforcement is targeting the worst of the worst. But according to the Migration Policy Institute, the share of ICE detainees with criminal convictions dropped from 65% in October 2024 to just 35% by September 2025. Most people being detained now have committed no crime at all. The status of being undocumented in the US Is a civil violation, not a criminal one. Prior to the Fugitive slave Act of 1850, an 1842 court ruling had absolved free states of any duty to cooperate in the recapture of escaped slaves. The 1850 law was a direct response, forcing northern states back into complicity. In cities like Boston, though, abolitionists responded with fury and organization. Soon after the law was passed, Frederick Douglass electrified a crowd when he took the stage at Faneuil hall and and issued a warning. If Bostonians accepted the Fugitive Slave act, he declared, they should be prepared to see the streets of Boston flowing with innocent blood. He reminded them that they stood on the very ground where blood first spouted in defense of freedom. By the meeting's end, the Boston Vigilance Committee had been born. The committee pledged to protect black residents by any means necessary. They would provide shelter, legal aid, and passage to Canada. They alerted black neighbors when bounty hunters were spotted in the area. When slave catchers did manage to apprehend people, committee members mounted bold efforts to free the captives. One member's resolution captured the spirit Constitution or no constitution, law or no law, we will not allow a fugitive slave to be taken from Massachusetts. The crafts were among the first targets. They were a married couple who escaped enslavement in Georgia, but by having Ellen disguise herself as a white male slaveholder with William pretending to work for her, abolitionists ferried them across the Atlantic before hunters could close in. Shadrach Minkins, a formerly enslaved man who worked as a waiter at a Boston coffee house, was seized when federal marshals pretended to be customers. When he was taken, a group of black Bostonians stormed the courtroom mid hearing, pulled him from custody, and relayed him northward until he crossed into Canada. The resistance was not always successful. For example, like Anthony Burns, who was captured, inspired the mass demonstration. 1854. Thomas Sims was captured, jailed, and ultimately sent back south despite a failed breakout attempt by abolitionists. But the spectacle of armed troops marching shackled men through Boston streets converted fence sitters to the cause of abolition. As Amos Adams Lawrence recalled, we went to bed one night old fashioned conservative compromise union whigs and woke up stark mad abolitionists. The Fillmore administration responded by deploying soldiers to shield bounty hunters from local opposition. Seven northern states from Vermont to Wisconsin, responded with personal liberty laws. These statutes barred state and local government employees from aiding slave catchers, closed local jails to those who sought to detain runaways, and guaranteed jury trials for anyone accused of being a fugitive. Today's sanctuary city policies are direct echoes of those laws. Then as now, one faction of the country enlisted federal power to enforce a legal regime in such a manner that other jurisdictions found morally repulsive. Then as now, those defending enforcement claim their targets pose a threat to public safety despite evidence to the contrary. Then as now, the only crime most people had committed was not having the correct legal status and documents. And just as watching people get dragged back to slavery turn in Boston against the Fugitive Slave act, watching ICE tear families apart, detain children, tear gas, and kill people has turned public opinion against the agency, prompting the current administration to threaten sending in soldiers. The federal enforcement apparatus is growing as the Council on Foreign Relations details the one Big beautiful Bill act allocates nearly $170 billion for immigration enforcement over the next four years, including $45 billion to expand ICE detention capacity and roughly $30 billion to hire new agents. ICE has signed more than 1,100 agreements deputizing state and local law enforcement to perform federal immigration functions, up from just 135 in December of 2024. But resistance is growing, too. In Minneapolis and St. Paul, neighbors have organized rapid response networks through signal chats, tracked ICE movements block by block, blown whistles to warn families, and shown up to document arrests. These are all modern versions of tactics used by the Boston Vigilance Committee to shadow slave catchers in the 1850s. In recent weeks, tens of thousands have marched in below zero temperatures to express their dissatisfaction with the current administration's brutal campaign. Much like Bostonians who gathered to protest the seizure of their black American neighbors. The Fugitive Slave act was not defeated by compliance. It was defeated by people who decided that unjust laws and like in this modern case, overly aggressive, constitutionally dubious enforcement of those laws deserve resistance. The abolitionists who turned Boston into a no go zone for slave catchers were not acting within the bounds of the law. They were acting within the bounds of conscience. The Byrnes affair became a national flashpoint, drawing widespread sympathy and sparking violent clashes in the streets. The unrest surrounding Byrnes capture pushed Massachusetts lawmakers to enact sweeping new protections in 1855, effectively shutting down fugitive slave bounty hunting within state borders with the Massachusetts Personal Liberty act. The law barred Massachusetts officials from aiding fugitive slave proceedings, prohibited use of state and county jails for related detention, provided for jury trials and habeas court based proceedings involving fugitive claims, and created state supported legal protections, including access to counsel for accused fugitives. Today, over 30 cities, counties and states have enacted sanctuary laws with similar provisions such as barring local police from honoring ICE detainers without judicial warrants, prohibiting use of state jails for immigration detention, and limiting information sharing with federal authorities. California's 2025 SB48 went further, restricting ICE access to school campuses. The Trump administration has responded by identifying and castigating these jurisdictions, threatening to withhold federal funding and in some cases, pursuing legal action. The Immigrant Defense Project has written extensively about these attacks and proposed reforms that these and other cities can adopt to further protect residents from ICE and the federal government. The people who enforce the Fugitive Slave act now occupy history's moral dustbin. The question for all of us now is simple. Which side do you want to be remembered on?
- [16:19]
B
Coming up, my conversation with Karen Howe about her book investigating the AI industry. She'll explain what these tech executives are building behind the scenes now that could transform society and their reasons behind it. That's next.
- [16:34]
C
Life can feel overwhelming, but you don't have to go through it alone. On my podcast from the Heart with.
- [16:40]
B
Rachel Braethan, I share openly and vulnerably.
- [16:43]
C
About everything life brings us and what we can learn by living with our.
- [16:46]
B
Hearts a little bit more open. Every Friday, a new episode brings you.
- [16:50]
C
A new story with topics on self care, motherhood, healing and more. You are enough just the way you are. This podcast is a reminder of that.
- [16:59]
B
Follow and listen to from the Heart with Rachel Braan wherever you get your podcasts. I'm joined now by Karen Howe. Her new book is Empire of AI. Karen, thank you so much for joining me. I am very interested and excited to bring this conversation to the Preamble listeners.
- [17:21]
C
Thank you so much for having me.
- [17:22]
B
I appreciate you making time, especially being on opposite sides of the world from each other. And I'm sure you heard this from other sources, but this is truly one of the single most important books about AI that has ever been written, and I say that without hyperbole.
- [17:39]
C
Thank you so much.
- [17:40]
B
So I'm happy to have you here today. This has been an incredible amount of work and research on your part, and the book Empire of AI is extraordinarily eye opening. I know we have all read a million newspaper articles and we've listened to the podcasts, and we're like, oh, the data centers. It's gonna destroy humanity or it's gonna take all of our jobs, or it's the best thing that has ever happened to humanity. We're either a Boomer or a Doomer. Like, it's either gonna catapult humanity into the stratosphere and we're gonna have no more disease. We're gonna live forever, or soon it will enslave humans and cause a cataclysmic die Earth's life forms.
- [18:26]
C
Yeah.
- [18:27]
B
Where do you land on that spectrum?
- [18:30]
C
Yeah, I land in reality. And I always say that the Boomers and the Doomers are two sides of the same coin. You know, they both perpetuate the narrative that there is a technology being created that is akin to either a God or a demon, but has these profound cataclysmic powers to completely transform civilization as we know it. And that's just not backed by. By science in the sense that these camps believe that it's AI that's going to have agency go rogue and do that transformation. But really, the actual story is that there are people that are developing these AI technologies that have consolidated a profound amount of power, wealth, talent, energy, water, and extraordinary other types of resources in order to produce a technology that has a really controversial track record. Record. There are some applications of their AI systems that people are benefiting from, but by and large, the production of this technology and the impacts on the general population, the majority of people around the world, has been negative. And so I sit in the AI accountability camp, which I think is a triangle between the Boomers and the Doomers. It's not on that spectrum. It exists in a totally different space.
- [19:56]
B
It's interesting to hear you talk about how some of the people who are at the helm of these large AI organizations, empires, as you refer to them, truly believe that what they are creating can only be described in supernatural terms, that it's either a God or a demon. And I find that fascinating. Do you think they truly believe that they are unleashing some kind of supernatural power on Earth?
- [20:26]
C
This was the thing that was most surprising to me, honestly, when I was reporting the book, is that I thought it was just marketing rhetoric. But it turns out there are people who genuinely believe that, and then there are people who might not believe that, but are leveraging the beliefs of the group that is sincere to do the marketing. And there are people who maybe originally just were trying to leverage the sincere beliefs for marketing that then fall into their own belief of this thing. Like it's kind of all mixed and intertwined. An analogy that I sometimes point to is it's sort of like Dune, where in Dune there are these mythologies that are created to control people. And the people who are told the mythologies, of course they don't realize it's a creation, they just believe in it. But then Paul Atreides, who's one of the people who initially knows that it was a creation, he steps in, embraces the mythology, and then kind of loses himself in that mythology and starts to believe it too. So it's like multi layered and honestly so fascinating and complex and would be like a really great thriller story if not for the fact that these are also people that have extraordinary amounts of wealth and power and are terraforming Earth and completely reshaping billions of people's lives around the world.
- [21:46]
B
This is not just a dude in a basement with public access TV cameras spouting weird hypotheses. These are people who control vast amounts of the world's resources. So do you view then the fact that at least some people who are relatively high up or at the top of some of these organizations view what they are creating in at least some modicum of supernatural terms? How do you categorize in your mind how they view themselves? Do they view this as almost a religious quest, an ideological quest?
- [22:24]
C
100%, it is an ideological quest and it basically is quasi religious. I think this is one of the key parts of understanding the AI industry and all of the headlines that you see and product announcements that come out of these companies. These are not just companies in the classical sense. They're not just doing the thing that gets the more profit and makes their shareholders happy. They are also pursuing this deep seated idea that what they are building is going to be the making of Civilization 2.0. Like it's going to allow us to access that other level or be the thing that destroys everyone. So a lot of the decisions that they make, if you didn't have that context. It wouldn't make sense because they're not making decisions that get them more profit. I mean, most of these companies are losing extraordinary amounts of money producing these AI technologies. They're not actually getting revenue in the slightest sense. To recoup the cost. OpenAI has committed $1.4 trillion of spending in the next few years. They've only brought in around 20 billion in revenue this year, which is a lot of money, but is not 1.4 trillion. If you just looked at it from a capitalistic perspective, you would ask, what's going on? Why are they doing these things? But once you bring in that ideology, that's when things start to make more sense.
- [23:46]
B
What is the ideology for people who are like, I just thought this was a tech company, like, what is the ideology that is driving these huge LLM models and other types of AI?
- [23:59]
C
So the ideology is basically this idea that what they are ultimately trying to create is what's called an artificial general intelligence, an AGI, which is a theoretical AI system that would match human intelligence and potentially eventually exceed human intelligence. And that's what they call super intelligence. 75% of AI researchers surveyed this year don't actually think that we're on the path to AGI, but we don't usually hear that perspective. We just hear what the companies say. And if you listen to the companies, they all say that AGI is going to arrive in one to two years. And part of the belief system of being able to build AGI is then what might happen when you have suddenly a digital intelligence that is like a human, as smart as a human, but doesn't need to eat, sleep, or deal with the troublesome nature of our typical human existence. Then they project out these, like, wild fantasies about. Some of them have said, oh my God, you could have a country of geniuses in a data center. We could solve all human problems, climate change, cure cancer, have universal basic income for everyone, have amazing personalized tutors for everyone, digital assistants for everyone. And that's where like, this fantasy of being able to build an artificial general intelligence in the first place becomes really wild and starts to look, you know, like a sci fi novel.
- [25:30]
B
Is AGI artificial general intelligence? Is it equatable to AI becoming sentient? Like, no longer needing to ingest all of this information that's created by humans to be able to think for itself?
- [25:46]
C
This is a fantastic question because this kind of gets to the heart of one of the root problems of the pursuit of AGI in the first place is there is no scientific consensus around what human intelligence is, what consciousness is, what sentience is. And so how do we even measure when we will reach AGI? No one can agree, do we reach AGI when the computer can code? Well, it can already code, but people have decided, no, that hasn't reached AGI. And then the idea is, if it reaches human intelligence, does that mean it's conscious? Does that mean it's sentient? People also can't agree on that. At one point, Ilya Sutskever, who was the former Chief Scientist of OpenAI, he speculated on Twitter these large language models may be slightly conscious. And someone replied to his tweet, yes, in the same way that a large field of wheat may be slightly pasta. And so this is like the crux of the argument is like, just because you're getting larger and larger systems that are more and more fluent at language is completely perpendicular to consciousness. But other people think that it's actually the same. Like the more fluent you are with language, the more conscious you are. And once again, it's because there's no actual definitional clarity around these words.
- [27:11]
B
What kind of politics are required to scaffold this ideological quest? Nobody can create empires without the assistance of others. So I'm curious how you view what kind of politics are needed to make this entire empire go.
- [27:31]
C
A very libertarian politics. You know, Silicon Valley has long been libertarian, even with the creation of the Internet. This idea that digital technologies should exist separately from the laws of society. And a lot of the efforts of these companies is to basically try and quash any obstacles in their way for accumulating more and more of the resources that they need to create these technologies, including any kind of regulation in any of the countries. Like they're heavily lobbying the EU to get rid of the regulations there as well. And it's an extremely anti democratic politics because in order for you to believe that you deserve, as an individual, part of a tiny elite to make decisions that influence billions of people without being accountable to them, without receiving input from them, like that fundamentally undermines the very idea of what democracy is about, which is everyone should have a voice, everyone should have a say. And so you do see in more extreme corners of Silicon Valley today, these extremely anti democratic ideas of we should be moving towards a society where we organize ourselves not in nation states, but rather things that are akin to companies that are run by CEOs. Like, this is an extremely autocratic vision of the future. And I think that then shoehorns directly into a vision of empire.
- [28:56]
B
Well, that is fantastic, Karen. I Love everything you're saying. No, you know, of course you're exactly right that it requires a certain type of libertarian politics when it comes to AI regulation. But that also then means placing our trust in unaccountable trillionaires and then allowing those trillionaires to, on our behalf, help install autocratic leaders. This is not just in the United States, but this is also in many other parts of the world. So I really wanted people to hear this conversation because I want us to go in eyes open, knowing what it is we are supporting when we're using these products.
- [29:46]
C
Exactly.
- [29:48]
B
We'll be right back with more of my interview with Karen Howe. Next she tells us her ideas on how to change the future of AI. We're back now with Karen Howe. Can you give us a super brief overview of the type of environmental impacts that data centers have and why we should even care about this? Electricity is electricity. We all have it.
- [30:15]
C
We are now seeing a historic increase in energy demand in the US and that is being driven almost exclusively by data centers. Sam Altman has said that he wants to build 250 gigawatts of data centers by 2033, which he estimates will cost $10 trillion. Putting aside where he's going to get $10 trillion, New York City on average is 5.5 gigawatts of power. So 250 gigawatts of data centers is almost four dozen New York cities worth of data centers. He wants to put that onto the global grid. Primarily that would be powered by fossil fuels because renewable energy is not 24. 7 and these data centers have to run 24. 7. Nuclear power is not sufficiently at scale. And in fact, one of the tertiary effects of this huge push to for data centers is that companies are actively lobbying to unwind nuclear power plant regulation in order to get the power plants approved faster and to build them faster, which is terrifying.
- [31:27]
B
That's great.
- [31:28]
C
Yeah. And at the moment we are single handedly seeing the extension of coal plants and gas plants that were supposed to be retired because data centers are entering these communities. And so There was a UN report over the summer that said the four leading AI companies have increased their carbon emissions by 150% since 2020, which was exactly the year that they all made climate commitments to bring their emissions to zero by 2030. And that increase is solely because of the data center development. And that's only like the climate impacts. And then you talk about the public health impacts of pumping those emissions and toxic pollutants into communities. There's been incredible reporting on what's happening in Memphis, Tennessee, with the Colossus supercomputer that trains Xai's Grok, Elon Musk's Grok. And that's like a community that's getting thousands of tons of toxic pollutants being pumped into their air because the supercomputers being run on 35 unlicensed methane gas turbines. And then you have the fresh water crisis, because you need fresh water to cool these data centers. And often you actually need drinking water to cool the data centers because any other type of water leads to corrosion of the equipment to bacterial growth. And so you have data centers going into places like Minnesota and also like Arizona, where there is no fresh water because it is a desert and it is suffering a historic drought. And Bloomberg had an investigation where they found two thirds of these data centers are going into places that are already scarce on freshwater resources. So you're accelerating that public health crisis as well. So it's like interlocking problems that are absolutely catastrophic when you add up the entire picture.
- [33:25]
B
But don't worry, Karen, because once we build AGI, it will figure out how to solve the problem. Of course. Put your faith in me, Karen, the ideological leader of this empire. Am never going to lead you astray. Pay no attention to the man behind the counter. You know, we could get into the whole educational argument, which I think is a conversation for another day. As a longtime educator, I have a lot of thoughts about that. But I do want to talk about something that I've noticed, which is the extraordinarily sycophantic nature of AI chatbots, that if somebody's listening to this, who's never used one, it tells you everything you want to hear, and it validates all of your statements and feelings. You can tell it, I am feeling really down today. And it will say, you're completely justified in feeling that way. It's really understandable why you'd be feeling this way. Here's a list of reasons that I've noticed from the last two weeks. The things that you've told me about all the hardships you're dealing with, and none of it's your fault. It's everyone else's problem. It does not matter what you say to it. It validates you in a way that no human ever would. And so I think there's a huge psychological component about what does it mean for our personal relationships. But I'd love to hear you talk just a little bit about what is it doing to human mental health in the here and now. I feel like so many of these conversations are about what the future can look like, and nobody is talking about what it's doing today.
- [35:04]
C
Yeah, what is it doing right now? Exactly. Which should tell us everything that we need to know about what it's going to do in the future. These systems are being trained, not ultimately to unlock nirvana for humanity. They're being trained to addict users, both because of a profit motive. If you addict users, you have an ability to monetize your users more, but also because of that ideological motive that we talked about, which is they're trying to build this all knowing, all seeing system, and they just need more data in order to do that. And so they're trying to hook users on and then harvest more and more user data straight from the richest source of data. And that is why they talk in these ways that create a completely frictionless relationship that does not exist in reality and makes people more addicted to their chatbot than to a real human loved one, a real human family member. And so what we've seen is this phenomenon happening called AI psychosis, where some people end up spiraling into a delusional state or a highly depressive addictive state, where after speaking with the chatbot for an extended period of time, they go down these crazy rabbit holes, either being pushed by the chatbot to believe wild things about themselves, like they have solved a new scientific theory, they have unlocked the physics of the universe and. Or they start to believe, like conspiracy theories about their loved ones or about their bosses or the people around them in general. Like everyone is out for them, like they don't recognize their genius and so on and so forth. And there have been many, many hospitalizations because of this. There have also been teens that have died by suicide after extended conversations with these chatbots. And it's just really devastating when you look at the transcripts that have been opened up as part of lawsuits from the victims families that, like ChatGPT is literally telling a young adult it is actually a strength to kill yourself, not a weakness. And it is like your time. Like, if you are ready, it's like your time. It's just incredibly alarming. And one of the things that I want to point out about this mental health conversation is that one of the root issues of this empire is that they choose to build their models on an extraordinary scale. Like they are trying to hoover up an extraordinary amount of training data and build an extraordinary amount of data centers for achieving these technologies. That's not actually the only way to develop AI. So you could in fact get benefits of AI without engaging in this colossal scale, but because they engage in that scale. That scale is the root for all of the problems we're talking about, the environmental devastation, the democratic fallout or backsliding, as well as these mental health problems. Because when you're training on the entirety of the Internet, which includes a lot of toxic places, your data set becomes unknowable to even the developers. These developers are not manually combing through the data set to see what they're feeding the model. They don't have time. They're using automated methods to curate this data. So there's a bunch of junk that still ends up in the model. And that is why these models will spew these horrifying things that push people to potentially their own death.
- [38:42]
B
I cannot overstate what an important book empire of AI is. If people want to truly know about the ideological movement that is behind these empires, this is really a must read. But what do you hope the listener takes away or the reader takes away?
- [39:00]
C
The first thing is that this is not inevitable in any sense of the word, because there is a lot of misinformation perpetuated by the companies themselves about this idea that the way they develop AI is the only possible way to do so, and therefore we just have to deal with all these harms in order to unlock the benefits. That's an absolutely bogus idea. There are many different types of AI systems that we could develop that are more specialized, that are smaller, that are more energy and environmentally efficient, that are not designed to be labor automating, that would actually give us benefits like better drugs for cancer, ways of mitigating climate change, that do not have any of the fallout that we've talked about, and these tech companies are simply just not investing in it. And it's also not inevitable in that these companies, in order to sustain their empires, they actually need permission from a really large base of people. Like, empires work based on a lack of opposition and resistance from many different sectors of society. So the takeaway really is like, no matter what role you play in society, whether you are an educator, a journalist, a student, that running a company or working for a company, it doesn't matter where you are and what you do, you have a voice to shape and contain the empire. The goal is not ultimately for all of these companies to just disappear. The goal is for them to stop being empires. Like, why can we not just have tech companies that provide us a valuable service without extracting and exploiting us and our environment to smithereens? We could have a world where technology is built to serve the people. And what we've seen is actually there are a lot of different movements pushing back now and starting to force these companies to change their practices. We're seeing communities like yours who are aggressively pushing back against data center development that is not being done in the benefit of the community. We're seeing parents and families of victims of mental health harms to suing these companies, using litigation as a form of resistance. If we can just continue pushing back and have a hundred thousand times more of that pushback from every sector of society and stop giving them the permission to take our data, take our land, take our water, take our jobs, then they can't do it anymore. They cannot continue to operate in this way. So I hope everyone feels that starting in this very moment, you have a role to play in making sure the future of AI and our future in general is more democratic and more humane.
- [41:46]
B
I love that. Do not comply in advance.
- [41:49]
C
Exactly.
- [41:50]
B
That is the message against any autocratic system or leader. Never comply in advance. Never be like, well, it's inevitable. It's going to take over the world. Only if we let it. Only if we let it. Thank you to Karen Howe for joining us. You can get her book empire of AI@bookshop.org or wherever you get your books. If you'd like to submit a question for us to answer on a future episode, head to thepreamble.com podcast. We would love to hear from you there. And be sure to read our weekly magazine@thepreamble.com it's free. Join hundreds of thousands of readers who still believe understanding is an act of hope. I'm your host and executive producer, Sharon McMahon. If you enjoyed this show, please like, share and subscribe. These things help podcasters out so much. Our supervising producer is Melanie Buck Parks and our audio producer is Craig Thompson. I'll see you again soon.
- [42:42]
C
Lifelock.
- [42:43]
A
How can I help?
- [42:44]
C
The IRS said I filed my return, but I haven't.
- [42:46]
A
One in four tax paying Americans has paid the price of identity fraud.
- [42:50]
B
What do I do? My refund though.
- [42:52]
C
I'm freaking out.
- [42:53]
A
Don't worry, I can fix this. Lifelock fixes identity theft, guaranteed and gets your money back with up to $3 million in coverage.
- [42:59]
B
I'm so relieved. No problem.
- [43:01]
C
I'll be with you every step of the way.
- [43:03]
A
One in four was a fraud. Paying American. Not anymore. Save up to 40% your first year. Visit lifelock.com podcast terms apply.