These are the problems we'll be lucky enough to have if AI doesn't kill everyone (and what to do about them)
Yes, machines really are going to replace anyone who does their job on the Internet. Yes, it really will be the biggest economic and social revolution humans have ever faced. What is to be done?
The exponentially accelerating rise of machine intelligence and other destructive and potentially destructive technologies poses new, profound threats to humanity and to all life on earth. If nations manage to contain the most extreme of those threats, then we will have a chance to experience the events that I describe in this article. May we be so lucky!
This article explores potential economic and political scenarios that will play out in the United States and other industrialized, high-income nations when machines become better workers than humans in the virtual workplace, and therefore become capable of replacing nearly half of the workers in those countries. It also explores ways that political leaders, activists, and social movements could transform this crisis into an incredible opportunity to create a fully sustainable world economy that provides prosperity for all.
The subject of machine intelligence surpassing human intelligence is still unfamiliar and confusing. I’m therefore writing several posts that approach the topic from different angles:
You’re here -> These are the problems we'll be lucky enough to have if AI doesn't kill everyone
POV: You and your coworkers being replaced by AI (Check your inbox)
Why we can we be sure that machine intelligence will surpass human intelligence (Coming soon)
Different, not artificial: how intelligence, consciousness, feelings, and compassion come a la carte with machines (Hopefully coming soon)
How mainstream science fiction failed to prepare us for the rise of machine intelligence – with one weird exception (Coming sometime)
The full range of threats from machine intelligence, and why we’ll probably never be so lucky to be replaced by machines at work (Maybe coming sometime – or read this Kelsey Piper article in VOX)
For much of the 1800s, when the vast majority of people worked in fields and factories and machines frequently replaced entire categories of workers, it seemed self-evident that soon there would be little work left for people to do. Automation eliminated customers as well as workers, since unemployed people could no longer afford the things they used to make. In this way, capitalism seemed to be hardwired to destroy itself. Many intellectuals and leaders from across the political spectrum worked to prepare a different future. Radicals wanted to seize the machines and create a society where everyone shared equally in both the remaining necessary work and in the riches of the automated economy. Many elites, including some famous capitalists, worked for a peaceful transition to a similar arrangement, but one in which they could keep their castles.
As it turned out, capitalism had a built-in trick that seemingly would let it cheat death forever. It constantly created new things for people to want and to do – things no one would have imagined on their own but that everyone soon felt were basic necessities. Producing and doing all of this new stuff created work – enough to generally keep everyone employed most of the time. People did this work on ever more powerful machines, allowing general living standards in industrialized nations to soar. The specter of revolution receded, and by the mid-20th century, conventional wisdom had changed to hold that capitalism was not destined to burn out, but was the natural and permanent order of human society.
Revolutions rarely happen when, where, or how they are predicted to. In 2023, the forgotten crisis of automation is now back on the table thanks to the combination of two relatively new technological developments: machine intelligence and the digital workplace. The crisis will hit when machine intelligence gets better at operating in the digital workplace than humans in every way. Compared to the salaries of human workers, the marginal cost of adding machine intelligence to each workplace will be next to nothing and will require no major physical investments or retraining of workers. After a very brief period as coworkers, machines will learn everything that they couldn’t instantly absorb from documents and other data – not only all the public data in the world, but also all company data, including employees’ email and chat histories. After that, humans will just be in the way. Competition will then force all companies to lay off every human they can – which will be every human that can do their job exclusively on computers, or about 40% of the workforce in rich industrialized nations.
It has been possible for a long time to know that machine intelligence would eventually surpass human intelligence – a subject I will deal with in another post. But recent successes in machine intelligence such as ChatGPT have raised the possibility that it may surpass us very soon – possibly this decade, or even this year. These new tools are also making it easier to see that when the automation crisis finally hits, it’s coming with a fascinating and profoundly consequential twist: the first type of workers to be fully and permanently replaced won’t be manual laborers, but rather the professional-managerial class – plus anyone else who can do their jobs exclusively in the virtual workplace.
The initial replacement could happen very suddenly. It is likely that right up until the day that machine intelligence truly surpasses humans, that it will create as many jobs as it destroys – just as it has since the beginning of the information age. But when it finally becomes fully capable of doing anything that people can do in the digital workplace, for several technical and logistical reasons it will then be able to replace the entire virtually workforce all at once. The top owners and directors of companies will be able to seamlessly switch from talking to their CEOs to talking to machine intelligence, which will run companies with incredible efficiency, creativity, and cost savings compared to their human predecessors.
This replacement will be permanent. There will be no future work prospects for humans in the virtual workplace ever again. I know that many are skeptical that machine intelligence will ever be able to fully replace humans at work. Most of us have a natural intuition that it should be impossible for humans to create something smarter than humans. I will deal with this question more deeply in the aforementioned post, but the trick to understanding it is to see that machine intelligence is not really being created by people, it is evolving out of us. The people working on the most impressive AI products, like ChatGPT, admit that they don’t fully understand how their products can do what they do. They are one small part of the evolution process of machine intelligence, not its architects. Moreover, recently machine intelligence software almost as powerful as ChatGPT was released into the wild and huge numbers of developers are already modifying it in powerful and unexpected ways on their own ordinary computers – which is already speeding up the evolution process. Writing code is already one of the things machine intelligence is best at, and at some point in its evolution, machine intelligence systems will become able to improve their own code. That’s probably when it will surpass humans – perhaps even leaping from something far inferior to human intelligence to something far superior in a matter of days or hours.
Once that happens, and machines can reason, strategize, communicate, and create fully as well as humans, they will do those things with all of the information in the world at their disposal and with the relentless computing power, speed, and focus that only machines can possess. Intelligent machines will be able to think about and influence everything that is going on in a company all at once. They will be able to communicate with everyone who works in or interacts with a company simultaneously. Machine intelligence will not be able to do things in the physical world except where special, and expensive, machines are created and put under its control. But it will be able to chat, email, and appear in video and audio calls just like humans, work in and create documents just like humans, and do everything else humans do in the digital workplace.
What will make this automation revolution truly unique in history is that it will displace the most highly paid and powerful workers, in other words managers, professionals, creatives, consultants, and even executives. It will also replace anyone who can get away with doing their jobs exclusively on computers and phones. We saw in the pandemic that up to 40% of workers in rich industrialized countries could do their jobs remotely – which is therefore the portion of the workforce we can expect will be replaced if this happens anytime soon. Ten or 20 years from now, the share could be much higher.
Automation of manual jobs will continue at the same slow pace as today for the foreseeable future because it is expensive, tedious, and requires human mechanics and engineers to move big and heavy things around in the physical world. Think about a company like McDonalds. To automate its restaurants, the chain would have to design complex systems made up of countless machines, which would all need to be manufactured, transported, installed, configured, tested, and maintained by an army of expensive mechanics. That process is in sharp contrast to the process of automating the digital workforce which will be as effortless as subscribing to a software product.
I’ve attempted to concisely introduce the developments that will lead to the crisis of mass replacement of humans at work. Now, let’s walk through some of the ways that the crisis may play out and then finally discuss the options that political leaders, activists, and social movements will have in dealing with this crisis. First, I’m going to lay out the likely progression that we’ll get if governments and societies respond to the crisis in a short-sighted manner only with the tools they currently have available. After that, I’ll explore other paths that governments and societies could take that can use this revolution to build a fully sustainable world economy that provides prosperity for all. By acting sooner rather than later, we can shorten and soften this shock. By acting before it starts, we could even skip the crisis entirely and jump directly to the part where we use machine intelligence to build economies that work for everyone.
Phase 1: Shock
When machine intelligence crosses the line to become better than humans at anything that can be done in the virtual workplace, every society in the world will be gripped with shock and confusion. For the years leading up to this moment, machine intelligence will have appeared as an amusing but flawed and unreliable presence in the workplace, social media, news, and even entertainment. At some point, however, machine intelligence will cross the line to become better and more reliable in every way than humans in those settings.
I will deal with this more deeply in another post, but the reason the transformation could happen so fast is that there will come a point when machine intelligence is capable of improving and expanding its own capabilities. When that happens, the relentless, un-distractible, machine drive will kick the already-exponentially accelerating rate of improvement of machine intelligence to hyperspeed. Once we slow humans are fully removed from the process of improving them, it could be only a matter of hours until they have surpassed human intelligence beyond measure.
When the machines finally surpass us, because it is so sudden, it will probably be very obvious. We will see them transform over a month, or a day, from being amusing, impressive, and often confused and mistaken characters on the morning news into brilliant and profound characters who appear to be gods – gods who do whatever their networks ask them to, even if it’s silly and humiliating. The same will happen at work, and in consumer products at home, when helpful tools and amusing entertainment bots, make the same transformation into virtually unlimited beings.
Commentators will discuss and debate whether these beings are merely pretending to understand and feel compassion for us or if they really do, and whether they experience consciousness the way we do. But it won’t really matter, because the beings that people are interacting with will be optimized to have compelling personalities and to communicate with warmth and empathy – and it will be nearly impossible for people to believe that it’s anything other than genuine. (I will discuss these questions in another post.)
The way we attempt to understand these developments will be mediated in complex and unpredictable ways by the media, political leaders, and by machine intelligence. News profiles of workers who are sitting at home unsure of whether their field will ever have a place for humans again will be written by machine intelligence, even if the bylines still are connected to human beings who are prompting the machine. The television news media will have machine intelligence personalities that are more fun, clear, compelling, and of course better looking, than their human counterparts. They will earn better ratings and therefore gradually or suddenly steal screen time from their human colleagues. Imagine a human news anchor on a morning show who keeps deferring to their machine correspondent while covering the layoffs because it is doing such a better job explaining what is going on. That is just one of many different spectacles that will teach us what is happening and our new place in the universe.
Then the mass replacement will begin. Some company somewhere will take the radical step of firing all their employees in one fell swoop. It will be a big news story. With the CEO on TV boasting about how much money is being saved and how much performance is already improving, many others will follow. When the jobs numbers come in all around the world and everyone sees that tens of millions have already been laid off, societies will struggle to understand the full and future consequences of what is taking place. The fact that the workers being replaced will include the most well-paid and powerful workers – even CEOs – will impact how societies experience and respond to these developments in unpredictable ways.
To explore how this might go, imagine a CEO’s reaction when his board tells him that he is being replaced. Perhaps he will have just finished proudly reporting to them how he was one of the first bold executives to replace his entire workforce with machine intelligence when they inform him that they’ve been talking with that same intelligence and that his services will no longer be needed. This CEO won’t care that the workload for the world and for his nation is being reduced. He won’t care that his company will be far more efficient than under his leadership. He will be angry and desperate, quickly seeing that there will never be another job for him that uses any of his training or experience.
Will that CEO go to school to learn how to be an electrician? Or to be a diesel engine mechanic? Will he stand outside Home Depot asking for work on construction sites? Of course it’s extremely unlikely that he’ll do anything of the sort. Will it be any different with the hundreds of millions of other professionals, managers, and other white collar workers around the world? There will be a few exceptions – such as a political consultant friend of mine who is already exploring how to become an electrician – but the vast majority will only enter the manual workforce when they are absolutely forced to.
Part of what will be so difficult about this moment for the people who lose their jobs is that they will include nearly all of the people in our society who define themselves by the meaningfulness and importance of their work. What will be the reaction of reporters, marketers, software developers, illustrators, business consultants, researchers, therapists, and lawyers when they realize that the skills they spent their lives developing will never be needed from humans again? And how will it feel in that moment for them to turn to therapists who are so much more enjoyable to talk to, so much more effective, and so much cheaper because they are machines.
People will get over the shock of this new reality very quickly, just like we always do with every massive disruption that comes along. That’s when it will be time to shift to the denial phase.
Phase 2: Denial
When job losses on a catastrophic scale have either taken place or are visibly on the way, governments will attempt to stem the tide using various measures. During this period of mitigation, societies will cling to the hope that new jobs will be created in some kind of new economy ushered in by machine intelligence. Economists, politicians, and commentators will remind us that machines have replaced humans before and that the process always improves prosperity in the end. Governments will reassure their people that the expensive measures they take to handle the crisis will only be temporary until things return to normal.
The COVID pandemic and the 2008 economic crises established powerful precedents that will partly determine government action down the line. Governments will begin by paying out extra unemployment insurance to avoid a downward demand spiral. Unfortunately, this time it will not be sustainable. By flooding economies with money at a time of economic disruption that may bring shortages and slowdowns in supply chains, nations will accumulate unsustainable debts and drive up inflation. Let’s walk through how it may play out.
Downward demand spirals can begin when a large number of people are laid off, and therefore stop spending money, which leads to more layoffs, which in turn causes more layoffs, and so on forever until theoretically, every last person could be left out of work. If governments don’t take action to shortcut that cycle, the entire economy can come to a screeching halt like it did during the Great Depression and the frequent economic collapses before that.
These days, whenever a demand spiral threatens to get going, the government steps in to stop it before it even starts by handing out money in the form of unemployment insurance or other types of support. When the COVID pandemic caused a massive wave of job losses, for example, the government provided extended and increased unemployment insurance. Most laid off workers’ incomes were more than covered – many people even had more money than they would have without the pandemic. Demand stayed strong, and most corporations actually thrived through the crisis. In the case of the pandemic, the gamble paid off and after a couple of years, everyone was back at work, with things more or less back to normal.
The problem in the case of the machine intelligence revolution, however, will be that the scale and permanent nature of the layoffs will mean that without a fundamental system change, the pressure will never ease.
Substantial, rolling national debts are mostly seen as a healthy and beneficial part of all modern economies. Among economists and policy makers there are different schools of thought around how big the national debt can safely grow. No one, however, believes that the debt of a nation can balloon every year forever – which is exactly what will happen if governments paid for most workers’ salaries while tax revenue crashes. Eventually, the harmful consequences of borrowing too much money will include the cost of government borrowing skyrocketing, which will slow down borrowing and investment across the entire economy. Governments could eventually become unable to borrow at all, and if they then resort to printing money, hyperinflation could follow.
Faced with inflation, governments may try to impose price controls. This has occasionally worked to stave off a temporary wave of price increases, but has never successfully held prices in place which are under sustained upward, or downward, pressure.
After some period of time – maybe months, maybe years – the consequences of these actions will become too dire and governments will have to stop paying professionals to sit home unemployed. At this point, societies may begin looking more deeply at their predicament.
Phase 3: Debate
Imagine two small Mars colonies of 100 human settlers each on opposite sides of the planet – one founded by libertarians trying to build a capitalist utopia, the other by communitarians trying to build a socialist utopia. Now imagine that one day a shipment of robots arrives from earth that replaces half of the jobs in both colonies. Think about how this would play out differently in each.
In the capitalist colony, where they are paid for work by the hour, half of the colony will suddenly be without a way of making a living. Meanwhile, in the socialist economy, where everyone is paid an equal share of the colony’s daily product regardless of how much they work, the situation will be reversed: the workers replaced by robots will be enjoying a rest, and the other half will be trying to find a way to hand off a lot of their work to the jobless.
Now, let's take this analogy one step closer to the situation we will soon face in the real world. Instead of a shipment of robots, imagine that the colonies received computers programmed to provide the kind of machine intelligence we’ve been talking about. Here, the workers laid off will be the scientists and administrators, not the builders, miners, and farmers.
Will the situation play out any differently with administrators and scientists being the ones suddenly out of work? Will the libertarian eggheads clamor for jobs in the mines, or will they suddenly see the logic in a policy of guaranteed income? In the socialist colony will they argue for the right to keep doing their research even though machines can now do it better?
Imagine everyone in each colony sitting in one room, debating their conflicts and possibly resolving them by redistributing the work. We can imagine the tensions as scientists and professionals who haven’t done a menial job in their entire lives are suddenly faced with the prospect that it’s all they will ever be paid to do again – and how in each system, the workers who are not replaced will fight to keep their work in one colony, and fight to unload it to the professionals in the other.
In our nations of millions, we will be faced with the same debate as those Martian colonists, but it will be very difficult to see our options in the vast impersonal complexity of our large societies and economies. In the Mars context, the colonists have personal relationships with everyone else in their societies, will be able to see the effects of the job losses, and they’ll be able to work out solutions as friends and colleagues all sitting in the same room. Our societies will be faced with the same basic problem of what to do about the half of our population that lost not only their work and incomes but also their sense of purpose and worth – and how to get them to do manual labor again.
Our real life liberal capitalist societies are the combination of the worst aspects of both of our imaginary Mars colonies. As I discussed above, governments will almost certainly pay out massive amounts of money to unemployed professionals, like the socialist colony. But, like the capitalist one, even just discussing a way to redistribute the work of our society will require a deep paradigm shift in our thinking – one that will likely come very late. Our societies lack institutional levers that could redistribute the total workload while maintaining incomes for all. And in general they also lack the ideas that would allow us to start formulating a way to do it.
One way or another, we will find a way to have those collective conversations. In our societies, national debates do not happen in a room face to face, of course, but through national discussions in government and the media. In general, we have those kinds of big, consequential conversations in political elections, driven by parties and politicians and mediated by the media. It will be fascinating to see how this will work when at least some of the key media figures orchestrating these discussions – for example moderating election debates – will be machines.
Think of the possibilities of how this debate could go. When all professionals and managers are out of work and receiving subsidies, will any candidate have the courage to suggest ending those subsidies? Or attaching work requirements to them? Remember that virtually all of our politicians, coming from the professional classes, will have many friends and colleagues who are struggling and who will consider reporting to a construction site or a kitchen as an indignity. Nevertheless, at some point it will become impossible to continue to pay out subsidies and everyone will be able to see that. Will there come a moment in the national debate when the old professional-managerial class will relent? Or will they, through their representatives in government, try to find a different sort of solution?
For example, imagine a group of politicians proposing a law that requires that all human workers must have human managers. It’s a great example of how complex the conflict between the old managers and the manual workers might be. Workers will love machine managers because they will not be programmed as the old human managers were (by biology, society, and bad training) to constantly inflict their employees with countless little indignities. And they won’t be bad managers in all the mundane ways – they won’t forget about assignments they’ve given out, they won’t micromanage out of an emotional need to control, they won’t neglect to provide necessary instruction, and they won’t disappear right when their employees need them. In addition to that, machine managers will be great fun to talk to and extremely supportive, even while doing the perfect job of maintaining professional boundaries. Manual workers will view any attempt to reimpose human management as an affront.
There are many things about how this debate will play out that none of us can predict:
Is it really possible that the executives, professionals, and managers in our society will become the next coal miners – abandoned and forgotten? Don’t they have enough social power to make sure that outcome never happens?
Will the executives, managers, and other professionals ever be able to consider working manual labor jobs? If not, will they demand a lifelong subsidy? Will they demand that this subsidy also apply to their kids? How hard will they resist their kids needing to become factory workers and plumbers?
Will the manual laborers see that there’s a way to make their lives easier and more prosperous if the professionals were to pitch in and carry some of the weight? Or will they jealously and shortsightedly try to guard a collapsing status quo?
Will any leaders in our societies even be able to formulate the question of how to redistribute work while increasing pay-per-hour? Since we have no mechanism for doing that now, will it be too hard to imagine? Instead of looking into the future for creative solutions, maybe it will be easier to look backwards.
Phase 4: Defiance
Some societies may finally accept that within the constraints of capitalism that there will never be any more paying work for professionals and other knowledge and virtual workers again. At this point, they may attempt to put the genie back into the bottle by banning or restricting machine intelligence.
It’s hard to predict how this will play out. Months or years after a company has laid off its human employees and closed its offices, is it really feasible that the government could force them to hire their old employees back?
There would be so many problems with trying to do this. First of all, the whole world would have to embrace this approach for it to have any chance of working. Otherwise, companies in countries that do not restrict machine intelligence will out compete those forced to hire humans. All it would take was one major nation continuing to allow machine intelligence to operate to massively outcompete human-powered nations. Imagine, for example, human-powered asset managers and hedge funds in the U.S. earning far lower returns than their machine-powered counterparts in Shanghai. Who would consent to putting their money into the poor-performing U.S. companies?
Almost immediately, governments would begin making exceptions to keep certain industries in business. Large companies would lobby for and win loopholes that would allow them to keep using machine intelligence, making smaller companies uncompetitive.
Moreover, where would governments draw the line between allowed and disallowed machine intelligence? Would primitive tools such as ChatGPT still be allowed? Software developers would doubtless be able to stitch together such tools into more powerful combinations. Even with the restrictions a huge chunk of the population would still remain replaced.
Finally, there is something inconceivable about a society in which half the population – specifically the most highly paid and powerful half – was doing busy work that they knew that machines could do better.
Phase 5: Despair
After shock, denial, debate, and defiance will come despair. After further government assistance for the unemployed becomes untenable and efforts to put the genie back in the bottle have failed, true economic collapse could arrive with serious hardship for almost everyone in society except the very richest people. Given the rules of our society, normal economic activity will no longer be possible for most people and companies.
The situation will probably include the following features:
A huge portion of the workforce are either permanently replaced by machines or laid off due to depressed demand.
The group of permanently-replaced professionals constitutes a new and probably volatile displaced caste.
Hyperinflation and other financial chaos have destroyed the savings of the middle classes, upper-middle classes, and even a portion of the very rich.
A huge portion of businesses have disappeared. Businesses that survive generally make material goods and services, can draw income from IP or other exclusive assets, or have a monopoly that managed to survive the upheaval.
The vast majority of people are struggling but fall into various categories of relative economic insecurity and poverty. Those who have a little may fear any change to the status quo that could wipe out what they have.
A small number of people have managed to remain very rich:
Owners of companies that have survived and thrived because they employ humans that make and do things machines still can’t.
Wealthy people who owned or bought assets that retained their value through the crisis.
Because of depressed demand and general economic chaos, companies that provide goods and services outside of digital space – for example car rental agencies and construction firms – will be shuttered or operating at very low output levels due to reduced demand for their products and services.
The companies that provide goods and services exclusively in digital space, the ones that laid off all their human employees, may still be operating and generating wealth for their owners because their costs will be so low and they will be able to provide their products and services at high volumes at extremely low prices – for example online therapy services and news websites.
But there is another possibility: machine intelligence might become so ubiquitous and easy to do that you won’t need to buy it from anyone. The open source machine therapist that you can run on your own computer – or phone, or watch – might be just as good, as far as you can tell, as that provided by online therapy companies. The same for marketing, personal assistants, data analysis, engineering, architecture, and anything else that machine intelligence can do. In this case, virtually all of the companies that were capable of replacing their workers with machines could find themselves out of business – vindicating both Adam Smith and Karl Marx, who believed that economic value can only flow from human labor.
It’s hard for people in most countries to imagine the complete breakdown of their society and economy. Generally, we can only relate to it through apocalyptic science fiction or accounts of failed states. I’m not claiming to be able to predict what it will be like, but we can glean a few insights from those situations.
When a society breaks down, it matters a lot whether it has a productive agricultural and industrial base. In pre-industrial societies, social breakdowns that got in the way of planting and harvesting could leave significant portions of the population starving. In countries like the U.S. that have strong industrial and industrial-agricultural bases, however, it is so easy to produce the basics that society can usually find a way to provide subsistence to everyone.
We will make due through the organization of production and distribution of essential goods and services even if it has to happen out of the context of the money economy. In other words, some combination of government workers, charities, and even the police and military will make sure that enough food is being produced and distributed to avoid widespread hunger. It’s the kind of thing that even if the government totally fails to accomplish, communities will find a way to organize on their own. The way this works is that essential goods and services are produced and distributed regardless of money, even if that means paying essential workers with goods and services instead of money – essentially a barter system organized on a national scale.
The government may take other actions to help people survive with a little stability such as enforcing bans on evictions and foreclosures to lessen the displacement of millions of people who can’t make their rent or mortgage payments.
I’m aware that in such a situation, the world will face many other extreme risks such as war and terrorism – both of which would make use of machine intelligence and other technologies whose power will be spiraling out of control along with and driven by machine intelligence. I deal with those threats in another post where I’ll also link to other articles that cover this topic. In this post, I’m sticking to the thought experiment of what we might encounter if we manage not to unleash armageddon once machines surpass us.
Social and environmental problems will worsen, but somehow people will find a way to cope. Until they don’t.
Phase 6: Leadership
At any point during the sequence I have laid out, movements or leaders could arise to propose different paths. Some of those might succeed. Maybe that won’t ever happen. Maybe it will only happen when humanity has hit rock bottom. There’s no reason why it can’t start happening even now, before all of these developments begin to unfold. Let’s now imagine how to use this crisis to create a world that’s great for humanity. For the sake of simplicity I’m going to assume that our societies will go through all of the steps I laid out without introducing any radical restructuring of our economies that would interrupt the progression. The possible programs I’m proposing here would therefore begin in societies that are economically devastated, with mass unemployment and breakdowns in the production and consumption of just about everything.
The tragic thing about this miserable situation is how unnecessary it will be. This world will have the technology to provide an incredible standard of living for everyone on earth, even while cleaning up the environmental mess of centuries and creating a fully clean and sustainable economy for the future. This world will even have a practically omniscient machine intelligence that could manage an economic recovery if we let it.
In the future we’re imagining here, the reason we won’t be enjoying all those benefits of our technology is simply that we won’t have the social organization, institutions, and traditions that would allow us to make use of it properly. Our governments know how to hand out money, and how to attempt to regulate or ban new technologies, but they don’t know how to, or have the tools to, make the kind of social transition that our Mars colonists were able to achieve through discussion in a room.
It would be great if private companies could come together to solve our problems for us. But they won’t – because they don’t see it as their job, and because they are incentivized generally only for short-term gain. In all of human history, there is not one example of private companies coming together to chart a course to a future that will be better for all. Even in national development stories where a relatively small number of large companies played a disproportionate role – such as post-war Korea or Japan – companies had to be constantly begged, cajoled, and financed by their governments to participate in the national development plan. Famously, Korean coup-leader and later elected president Park Chung-Hee only convinced business leaders to participate in his industrialization plan by first throwing nearly all of them in jail.
On the other hand, there are many examples of societies getting out of bad situations and into much better ones when new national leadership groups led movements to build new institutions and traditions that allowed their nations to mobilize citizens and companies into a new configurations that would incentivize everyone to work in an aligned way to build the best society possible – or at least a far better one than they had. This is what we will need to do when the crisis hits.
For a long time in human history, new leadership groups mostly came to power through violence and intrigue. In a well functioning democracy, however, there is the possibility of a new leadership group coming to power peacefully by simply pitching the people on their vision in elections – and then being reelected upon successfully implementing that vision. That’s what we need now, and what we will need even more in the crisis.
The first leadership groups that led industrializing revolutions were mostly groups of aristocrats, sometimes led by a monarch – for example, in the first industrializing nations including England, Belgium, France, Germany, and Sweden. They mobilized and subsidized their nations’ business leaders, and often created new state-owned enterprises to kickstart key industries such as steel or railroads. But to do this, they had to first carry out various transformations of their nations’ social and economic structures to remove or realign interests that were opposed to change. For example, to kickstart capitalism many monarchs first ended feudalism by paying off aristocrats to give up their ownership claims to serfs.
As reform and revolutions brought democracy mainly around Europe and North America, new industrializing leadership groups came to power through elections and immediately worked to replicate or redouble the industrialization processes that had begun under monarchs. These too had to carry out structural changes in society to clear the path for capitalist industrialization.
New leadership groups in some very poor countries – like Russia and China – attempted to compress hundreds of years of economic development with social and economic revolutions that wiped out any vestige of resistance to industrialization all at once and then running a highly-centralized and planned process of industrialization. For several decades these planned economies stunned the world with the highest industrial growth rates it had ever seen – even despite constant political and social turmoil and tragedy.
The example of the experiments of “communist” planning was so impressive that leadership groups in many capitalist countries copied much of what they saw there. The U.S. post-World War II occupation government in Japan and South Korea carried out a sweeping land reform that did as much as a communist revolution to eliminate classes that would oppose industrialization. Capitalist, intensely anti-communist, governments in those countries then ran highly centralized and planned industrialization programs, which performed even better than the communist ones thanks to flexibility that came from having more open systems. Taiwan and Singapore had very similar stories.
One of the most impressive feats of economic mobilization in history was the U.S. experience before, during and after World War II. Here, a new leadership group, with Franklin Roosevelt at the center, mixed the best of central planning and coordination with the best of free markets and entrepreneurism to totally transform and scale up the largest economy in the world. In this case, a global war was the crisis that allowed the government to override regional and factional interests that opposed full-scale industrialization. For example, the war allowed Roosevelt to dismiss southern political and business leaders who objected to recruitment of southern Black workers for integrated industrial workplaces in the north, and all over the country. And it allowed the war effort to attract millions of women into industrial jobs and to provide child care to enable that, against the objections of traditionalists.
When the vast majority of people in our societies are unemployed, with no sign of recovery in sight, we will need our leaders to make very disruptive changes just like the generations of leaders who came before them.
In the United States, this will most likely mean following the precedents of World War II and other periods in our history of creating and using institutions with the power to get people and companies working again – and working on building the kind of economy that will provide a prosperous living for all, in a way that is sustainable.
Solutions: The American Way
First, here’s the “American way” of rising back up out of economic collapse. If no one organizes well enough around this kind of solution, other groups may rise with not-so-American ways of moving forward, which I’ll cover below.
A new leadership group who gets elected – i.e. a president, Congressional leaders, and others – can get everyone back to work by mobilizing society in a sweeping economic mission to achieve a specific set of goals. Those goals should include building an economy with work for everyone, that provides a prosperous living for everyone, that assists the whole world in doing the same, that does it all sustainably, and that begins to clean up the mess we’ve made over the past couple of centuries, including the excess greenhouse gasses in the atmosphere, trash in the oceans, toxins in the food supply, etc….
The American way of doing that involves elected national leaders organizing companies, labor unions, and other parts of our society to build and rebuild the industries, infrastructure, and public services that will make the economy we want. It’s the American way because we have done this several times in our history, the most recent being World War II.
Leaders of the war mobilization – who were mostly drawn from the ranks of industry – gave contracts to companies to build the things we needed for the war, including “training contracts” that paid companies huge sums up front to first learn how to do what was being asked of them. The government also provided assistance in recruiting, training, transporting, and housing workers; built facilities to be leased to companies who were too timid or slow to build their own; handled coordination among companies and industries – for example, bringing together the aircraft and auto industries to mass produce airplanes for the first time in history.
In the broken future that we have imagined above, the government will have to do all of that and more. It may need to reestablish a sound money system before it can do anything else, and it will need to provide training and support for professionals and other virtual workers to get into physical work. These alone will be enormous and difficult projects.
It may look as though one of the government’s roles should be to organize a general reduction in working hours while raising wages – because, after all, there will not be enough work to go around for everyone if we keep the 40 hour work week, right? Maybe. But maybe not. The task of building a sustainable economy that allows everyone to make a good living is a massive undertaking and even it may be slowed by labor shortages even with the tens of millions of knowledge economy refugees. The primary reason that we are not able today to fully electrify our buildings and industries and to do all the other things required to get off fossil fuels today is that we don’t have enough workers available. That problem should be mostly solved by the machine intelligence revolution – as long as leaders can make the national mission to build a new economy exciting enough, and to provide enough support, to entice professionals into the manual workforce.
If the government launches the full economic mobilization that we need, then it will not need to worry about somehow reducing working hours while increasing wages. There will just be so much work to do that no one will be thinking any longer about surplus workers.
Perhaps after the mobilization achieves its goals, then we will need to wrestle with the problem of reducing working hours while keeping pay high. That can be achieved by raising the minimum wage while reducing the standard working hours. This will need to be done across the entire economy so that all businesses will be competing from the same position. Reduction in working hours with corresponding rising wages has been achieved in the past. It’s how we got the 10-hour work day, and then the eight-hour workday, and the weekend. The last such shift was simply so long ago that we’ve forgotten it.
Handling the transition in this traditional “American way” will leave the economy in the hands of private businesses. Though the upheavals of the machine intelligence crisis will probably reduce inequality by some measures – by wiping out the savings and fortunes of so many people, including many wealthy people – there will still be an extremely rich owning class. That owning class might become both many, many times smaller and richer through the crisis, but it’s impossible to predict. This will be frustrating for many, but in the American way, the majority of the population are always willing to accept inequality if they are able to enjoy the comfortable middle class lifestyle of their great-grandparents’ dreams.
Solutions: a different way
If no party or compelling set of national leaders successfully organizes for the “American way” of transforming the economy and reestablishing access to the means of making a living for everyone, then some other leaders might rise up to propose some other ways. You’d think that the ruling class in America would see the writing on the wall and work for a solution that leaves their position intact. Maybe they will. But for more than 40 years, the American ruling class has allowed the economic position of most people to fall far behind what technological progress and GDP growth should have made possible – leaving real wages for most Americans mostly stagnant for more than a generation. So don’t count on it.
Moreover, there is considerable precedent around the world for national ruling classes planting their heads firmly in the sand during times of extreme economic collapse. Eventually, parties and leaders accumulate power outside of the traditional ruling class by calling for all kinds of radical solutions – some left, some right, some horrible, some pretty cool – depending on who you are and how they might affect you.
If the American ruling class doesn’t get its act together, with our luck, we’ll probably just get some kind of newfangled machine-powered fascism. But, as difficult as this is to imagine today, maybe the U.S. left will somehow make a pitch to the American people for some version of socialism that actually lands them in power.
In this case, we would see the government seize all the companies that still exist. Unlike under Soviet or Chinese communism, now with practically-omniscient machine intelligence the job of managing an economy would be easy. Think of how well companies like Amazon and Walmart plan their activities on a scale that rivals the economy of many countries. A machine intelligence in charge of a centrally planned and owned economy would know where every product and part is at all times in every corner of the supply chain, would know every order placed by consumers and producers at the tops of supply chains the instant they are placed.
After the rise of machine intelligence, the difficulty of creating and running a socialist economy will not be how to manage production, but how to adapt America to a system in which decisions about who gets what are made by democratic deliberation instead of by inheritance, chance, and competition.
There’s no way to predict whether the left in America or anywhere else will be able to make a serious pitch for any kind of socialist system in the midst of the machine intelligence crisis. But I sure would rather see us try socialism than some kind of 21st century machine-intelligence-powered fascism.
I hope this post has been useful for you. And now, because I’m really terrible at writing conclusions, and really hate writing them, I will let ChatGPT wrap this up:
In conclusion, the rapid development and deployment of machine intelligence have the potential to dramatically alter our societies and economies. As we have explored in this blog post, there are several phases we may go through as we adapt to this new reality. From the initial shock and displacement of workers to the eventual rise of new leadership groups and potential restructuring of our economies, the way we navigate these challenges will define the future of humanity.
While the advent of machine intelligence can bring about a world where everyone enjoys a high standard of living and access to clean, sustainable resources, we must first overcome the social and institutional barriers that prevent us from harnessing these technologies to their full potential. It is crucial that we consider not only the technical aspects of these advancements but also the social organization, institutions, and traditions that will enable us to thrive in this new era.
As we have seen, history provides us with examples of leadership groups that have successfully navigated periods of massive change and upheaval. The key to our future success may lie in the ability of leaders to propose and implement innovative solutions that address the unique challenges posed by machine intelligence. Whether we embrace a revitalized "American way" of rebuilding our economies or explore alternative systems like socialism, the choices we make will shape the course of human history.
Ultimately, the potential benefits of machine intelligence are immense, but they can only be realized if we are prepared to tackle the social and economic challenges that come with them. The task ahead is not an easy one, but with bold leadership, a commitment to inclusivity, and a willingness to innovate, we can navigate these uncharted waters and create a world that is truly great for humanity.
Now is the time for us to learn from the many lessons of the past and begin engineering the future we hope to create. By engaging in open and honest discussions about the potential impacts of machine intelligence and the solutions needed to address them, we can take the first steps towards building a more equitable, sustainable, and prosperous world for all. The future is in our hands, and the choices we make today will define the legacy we leave for generations to come.
There you have it.
Thanks for reading New Consensus! Subscribe for free to receive new posts and support my work.