AI and Oil Wells: Will AI Destroy the Middle Class and Democracy?
The champions of AI claim it will create a new era of prosperity. It could destroy the middle class and democracy instead.
The cheerleaders of AI like to claim it will create a bright new era of plenty. They claim AI will allow firms to produce more things cheaply, producing a bounty that all humanity will enjoy. It will open access to expensive services like accounting, law, and medicine. It will empower people with good ideas, creativity, and drive to launch new companies. A whirlwind of AI agents will do work humans currently do, leaving us free to do the things we love while enjoying the prosperity AI effortlessly provides.
There’s no logical reason AI couldn’t in fact create such a utopia, but I doubt it. I worry it could just as likely throw us into a dark age of deprivation and oppression. This isn’t because I think AI will become conscious and super-intelligent and then oppress or kill us—although, honestly, that’s a concerning possibility. It’s for the same reason it’s a disaster when a small and poor country discovers oil.
THE PROBLEM WITH AN OIL WELL
It’s a counterintuitive fact of international development that among the worst things that can happen to a small and poor nation is to discover a valuable natural resource like oil. You would think finding oil or diamonds in a poor country would be a godsend. Since leaders now have something valuable to trade for hard money, they can use the windfall to build new infrastructure, roads, and hospitals. They can attract foreign investment to build up higher-tech industries. They can invest in healthcare and education. They can make their people happier, and their nation stronger and more powerful.
While rulers absolutely could do all this, what almost always happens is the opposite.
In the real world, what usually happens when a small and poor nation discovers a valuable natural resource is it becomes a horrible autocracy. The ruling class uses the wealth to fly about on private jets and buy expensive foreign goods. Instead of investing in infrastructure, or building a strong, prosperous, and educated middle class, the nation splashes money on a few showy projects. Rulers zealously guard their power and treat any challenge to their authority with brutality. The lives of ordinary people, neglected in squalor, don’t get better. They get worse.
When you think about it, the reasons this usually happens almost are inevitable.
It only takes a relatively small number of people to extract the value from an oil well or diamond mine. You can even sign a contract with a large Western firm and be done with it. The wealth, however, is sufficient to buy off everyone you need to solidify your power. In a small nation, you only need to buy off a handful of key people—a few generals, the head of the national police, some figures in the media, maybe leaders in the church, and some local leaders like the heads of tribes or regional strongmen. If you keep this handful loyal, you can maintain a near-unchallengeable lock on power—and now you have plenty of money from which to give each of them a generous cut.
When you have an oil well, why fund roads or infrastructure? Why provide healthcare or education? Why create a prosperous and educated middle class that will only build up expectations and new blocs of power, making your situation weaker? Well-meaning and naïve do-gooders say you should do it anyway. It’s not just the right thing to do, but will benefit you and your nation over the long term. Investing makes your country smarter, healthier, happier, and richer, and ultimately more powerful. It creates new industries and thus more sources of wealth. It increases your influence and hard power. However, what the do-gooders fail to understand is that when leaders take this advice, they open themselves up to getting couped.
Resources spent on roads and hospitals are no longer going into the pockets of the key supporters you need to stay in power. If you won’t share it with them, inevitably someone less enlightened will. Someone more selfish and brutal can promise to take the money you’re spending on education and healthcare and give it to the generals, party officials, and police chiefs. Before long, an enlightened leader finds himself surrounded by gunmen in the presidential palace. Using any of the nation’s fruits to develop the country is just asking to get outbid.
Why then do modern democracies spend money on roads and hospitals and college educations? Are we more enlightened? Are we nicer people? The cynic would say it’s because the rulers of modern developed economies like America’s derive their wealth and status not from oil wells but financial schemes and stock markets—tiny slices of a productive economy the middle class creates. Our rulers panic when the middle class is struggling not because our rulers are more enlightened. They panic because an unproductive middle class means the people producing their wealth are producing less, which means the economy slows, which means markets plummet, which means the source of their own wealth and power gets destroyed. They provide health care and education because a happy and productive middle class performing office jobs is the source of their own wealth and power too. In other words, you’re their oil well. If you think this is too cynical, ask yourself why the powerful care a great deal about middle-class education but little about the homeless and disabled? The homeless and disabled aren’t productive workers, so no one is incentivized to care about what happens to them outside of a little charity.
That is why a diamond mine in a small and poor nation is such a nightmare. Of course, there are exceptions. This rule doesn’t seem to hold in large nations with vast resources, or in city states with tiny ruling classes. It’s impossible to control a massive nation like America, China, or Russia with just one major oil field. There are countless fields spread over vast distances, as well as copper mines, lithium mines, timber, iron, and more—although as Russia demonstrates, it’s not entirely impossible to control them all. Tiny city states, on the other hand, like the Gulf emirates can control an entire nation with a group so small they don’t need to be always looking over their shoulders and therefore can act with a bit more foresight. Nations developed before finding a resource, like Norway, also generally don’t follow the rule, since they developed their modern economy and powerful middle class before they found the oil.
The general rule, however, is you never want to live in a place in which a small group can control enough wealth to control the country without your help. Resources spent on you are resources a rival can promise the army—and while they need the army, they don’t need you.
IS AI LIKE AN OIL WELL?
If AI does what its proponents claim, it seems an awful lot like a diamond mine or oil well.
When people tell me how people could use AI to make a better world, I wonder if they’ve thought through how people will use AI to change the world. If AI is a magic money machine that doesn’t require a large and educated middle class to generate wealth, why does anyone still need you? When rulers no longer need you going to your 9 to 5, why go to the effort of building you nice housing in which to live? Why build roads and infrastructure to places they won’t go, which they will never use? Why provide you healthcare? If AI can cheaply generate an entire economy’s wealth, it can buy all the generals, police chiefs, judges, and party bosses needed. A small group that controls it can make themselves kings, if not gods.
The wishcasting over the world AI could create reminds me of utopian arguments from the dawn of television. Optimists used to argue this amazing new invention would create a golden age, putting an education machine in every home. People could take classes, learn new languages, or immerse themselves in culture from their couch. They certainly could have—and some even tried. It turned out, however, most people wanted to watch sitcoms and game shows, and eventually shows in which strangers emotionally damage one another like The Real Housewives.
Tech oligarchs also are naïve to think they’ll be the ones controlling this new technology just because they built it. Business people always think that. They will at first, of course, and might even cooperate with the ruling elite for a time. Soon enough, however, the people controlling the army, police, and courts will want rid of them, at which point they’ll find themselves like Jack Ma or Mikhail Khodorkovsky. Innovators and business people are dangerous loose cannons and competitors for power, and far less obedient and reliable than a flunky. They always end up like Daggett in The Dark Knight Rises, who thought funding Bane gave him control over Bane—only to learn as his neck cracked that Bane was always the one with power over him.
We’ve been obsessing over the technical risks of AI, while overlooking this just as troubling human risk. We may not know what AI will do because, whatever anyone confidently tells you, no one has any idea how AI entirely works. Does a calculating machine inevitably develop an identity, beliefs, goals, and sense of self? We don’t even know where human consciousness comes from, whether the brain is just a computer from which consciousness emerges, or whether consciousness comes from somewhere else. We do however have an idea what AI will tempt humans to do with such powerful new technology. It will tempt them to turn the middle class into the disposable class and destroy democracy.
This is the part of the essay where I’m obligated to say something hopeful and offer solutions. My best hope is the cheerleaders of AI are simply wrong. Perhaps AI never becomes everything they promise. Perhaps countless firms develop competing interchangeable AIs, along with scrappy individuals building their own AIs in basements, and no small group can ever control the technology. Perhaps AI never turns out to be a magic wealth machine at all, but just another useful tool that makes us all a little bit more productive like a stapler. However, what if they are right?
Then we must start planning now. We should be seriously discussing the social changes and institutions we need to ensure the human incentives for using this technology align with creating a society we want. Smart people must start talking not just about technological risks, but also equally dangerous human risks. If we’re not careful, the magic money machine won’t create the utopian fantasy we were promised. Like a diamond mine, it will create an autocracy leaving all of us outside the tiny inner circle behind.
What do you think of the rise of AI? Join the conversation in the comments.
I see the most likely human risk as more benign, but a likely precursor to the risks described here.
I fear that millions of ordinary people will look to AI as the Ultimate Talking Head, trusting it even more than they do our human elite to tell them what to do, precisely because it is not human … and continue to dismiss their own insights and common sense as irrelevant.
That makes them prey, for the Powers That Be.
This is interesting, but I think it likely misapprehends the nature of AI, and of wealth in general.
First, what is wealth? How do you define it? Probably, you have to define it in relation to material resources. The reason resources are a curse for undeveloped countries is only because there are in existence external developed economies that can produce the luxury goods a tinpot dictator wants; if his country were the only one in the world, then if he wanted luxuries he would have no choice but to develop his own country’s economy well enough to provide them. Diamonds aren’t worth much if there’s nothing you can buy with them that’s worth having.
The thing about AI is that it cannot, by itself, produce anything of material value. It’s a completely abstract device for symbolic manipulation. Yes, it certainly does threaten Musa Al-Gharbi’s “symbolic capitalists” in much the same way the Industrial Revolution threatened skilled craftsmen, but like the Industrial Revolution this will only produce more wealth insofar as it increases productivity, which will still require workers and that those workers can in turn afford to support the economy of mass consumption upon which the fortunes of the truly wealthy are built in modern market economies.
I foresee a period of political and economic turmoil not unlike what we experienced in the nineteenth century as the Industrial Revolution led to the evolution of mass politics, labor unions, Marxism, financial markets, etc. There will be winners and losers. There will be massive disruption. Societies could be politically destabilized, leading to revolutions as in 1848 and the establishment of new social contracts in their wakes. It may not be a happy time. But I do not foresee a new age of Feudalism or even a Dark Age; the wealthy elite will have to find ways to justify and legitimate their rule, as always, but totalitarianism is too likely to kill the golden goose: The mass consumer economy. A factory that can produce a thousand luxury BMWs a day is basically worthless if no one can afford to buy them. Likewise, internet companies are worthless if there is no one to advertise to or who can afford to use their services; the very network effects on which they depend obviously cannot exist without consumers to network, which don’t exist in sufficient numbers without an economy of mass consumption.
At the end of the day, masses of unemployed and miserable people are a revolution waiting to happen, which is why social democracies came to exist in the first place. I think this arrangement is too useful to be fully upended by AI; the terms might change (we could get a UBI for instance), but the basic structure of mass consumerism atop private property rights will not.