Cory Doctorow: Full Employment
I am an AI skeptic. I am baffled by anyone who isn’t.
I don’t see any path from continuous improvements to the (admittedly impressive) ”machine learning” field that leads to a general AI any more than I can see a path from continuous improvements in horse-breeding that leads to an internal combustion engine.
Not only am I an AI skeptic, I’m an automation-employment-crisis skeptic. That is, I believe that even if we were – by some impossible-to-imagine means – to produce a general AI tomorrow, we would still have 200-300 years of full employment for every human who wanted a job ahead of us.
I’m talking about climate change, of course.
Remediating climate change will involve unimaginably labor-intensive tasks, like relocating every coastal city in the world kilometers inland, building high-speed rail links to replace aviation links, caring for hundreds of millions of traumatized, displaced people, and treating runaway zoontoic and insectborne pandemics.
These tasks will absorb more than 100% of any labor freed up by automation. Every person whose job is obsolete because of automation will have ten jobs waiting for them, for the entire foreseeable future. This means that even if you indulge in a thought experiment in which a General AI emerges that starts doing stuff humans can do – sometimes better than any human could do them – it would not lead to technological unemployment.
Perhaps you think I’m dodging the question. If we’re willing to stipulate a fundamental breakthrough that produces an AI, what about a comparable geoengineering breakthrough? Maybe our (imaginary) AIs will be so smart that they’ll figure out how to change the Earth’s albedo.
Sorry, that’s not SF, it’s fantasy.
It is too late to halt the climate processes that will flood every coastal city, displace hundreds of millions of people, and sicken billions as pathogenbearing organisms seek new habitats where there is neither a resistance to them nor a predator to dampen the spread of their hosts. These processes will occur irrespective of geoengineering.
To understand why, consider just one factor: the heat we’ve sunk into the oceans. The seas won’t cool until the energy trapped in their depths is expended. Which means that, to a first approximation, the ice-caps are toast. I will speculate with you about GAIs all night long, but I’m not here for thought experiments in which we repeal the second law of thermodynamics. That’s not scenario-building, it’s wishful thinking.
The good news about all of this is that it reveals the locus of the problem with technological unemployment: it’s not a technological problem at all, it’s an economic one.
The pandemic crisis has taught us two critical things:
- Blind adherence to government austerity destroys capacity – it doesn’t build it.
California spent $200M in 2006 on a stockpile of mobile hospitals, tens of thousands of emergency beds, millions of N95 masks and thousands of ventilators.
In 2008, the state dismantled the stockpile to save the $5M/year maintenance budget (warehousing and keeping the ventilators’ batteries charged) so it could balance its budget during the financial crisis.
- Sovereign currency issuers do not experience cash shortfalls during crises – they experience capacity shortfalls.
Central bankers are getting carpal tunnel syndrome from typing extra zeros into their Treasury spreadsheets. The money supply is fine. That helicopter money is going to procure things that the private sector doesn’t want (the labor of people who are quarantined at home).
That money is mostly chasing the same goods that were available before the crisis (rent, groceries, and debt service) so it’s not crowding out other buyers and causing inflation (inflation occurs when more money chases the same goods, so buyers get into bidding wars that drive up prices – when the same quantity of money is chasing the same quantity of goods, there’s no inflation).
There’s one exception – one category of goods that is suddenly in much higher demand: health services, infrastructure, and supplies. We are severely short on hospitals, N95 masks, respirators, and trained medical staff (to say nothing of bleach wipes, gloves, gowns, and so on).
The Feds can make their money machine go brrr brrr brrr all night long, but that won’t make ventilators appear. The ventilators are not there. California threw ’em away 12 years ago to save money. California doesn’t need money; it needs ventilators. If, in 2008, the Fed had bailed out the states, replacing the money the private sector had just annihilated, California would have ventilators. Stockpiling medical supplies is smart planning. Stockpiling money when you are the source of money is… weird. It’s like Apple stockpiling iTunes gift cards or something. For you and me, money is crucial. For sovereign currency issuers, money is just integers. They’re not gonna run out of integers.
Which brings me back to technological unemployment.
The premise of automation-driven unemployment is that someday, the private sector will own so many robots that it can produce sufficient goods to meet all human needs without procuring any labor. Thus, no one will want labor, and so no one will earn the money to buy the things the robots are making.
Should that day come to pass, the money-issuer – the central bank – can procure all that idle labor, without creating inflation. If the private sector doesn’t want labor, then there’s no bidding war between companies and a federal jobs program or a universal basic income, so creating money to buy people’s labor won’t drive up the price of that labor.
What could the government ask us to do in exchange for all that money? How about doing the hard work to cope with the centuries of climate emergencies we definitely, unavoidably face?
Look: when the pandemic crisis is over, 30% of the world will either be unemployed or working for governments.
This isn’t after an Artificial General Intelligence Singularity in the distant future.
It’s next year.
Governments will get to choose between unemployment or government job creation. Governments that choose unemployment – that leave 30% of their population out of work – will collapse. Possibly their nations will collapse with them. Nations with 30% unemployment don’t function. More importantly, nations with 30% unemployment can’t do the capacity-creation work that will make them resilient to the next crisis, leading to a civilization-threatening death spiral of collapse-disaster-collapse-disaster.
One side effect of this is that it thoroughly forecloses on the dream of a General AI. Societies whose primary industry is digging through rubble for canned goods while drinking your own urine to survive do not make fundamental AI breakthroughs.
The 30% unemployment ”solution” solves nothing, but what about a jobs guarantee?
The 30% who are working under federal jobs programs will have working lives completely decoupled from the workings of the market. The movements of markets – particularly financial markets – will be irrelevant to everyone except for a group of weird, chart-watching, twitchy nerds who fulfill the boring job of capital allocation to an increasingly irrelevant section of the economy.
Banking will once again be a safe sinecure for the unambitious, none-too-bright children of wealthy families, the kind of people who want four weeks paid vacation a year, to clock out every day at 4:45 p.m., and never want to work weekends.
Can we imagine a jobs guarantee? You might find it hard to believe that we will find the political will to simply procure the labor of people who have no jobs. Even in the face of a climate crisis that desperately needs that labor to prevent the destruction of our cities, our civilization and our species, it’s hard to imagine that kind of ambitious program.
But honestly, if we’re going to dream, let’s dream big: I can see a path from here to a federal jobs guarantee. I don’t see a path from here to a General AI. If we’re going to imagine a General AI future, let’s imagine that one of the changes we make along the way is to abandon the toxic austerity worship of the neoliberal era. If we can manage that one simple trick – the trick of adding zeroes to a spreadsheet in a central bank computer – then we have no automation employment crisis.
Which is really good news, because that will free us up to address the actual crisis that we face.
Keynes once proposed that we could jump-start an economy by paying half the unemployed people to dig holes and the other half to fill them in.
No one’s really tried that experiment, but we did just spend 150 years subsidizing our ancestors to dig hydrocarbons out of the ground. Now we’ll spend 200-300 years subsidizing our descendants to put them back in there.
Cory Doctorow is the author of Walkaway, Little Brother, and Information Doesn’t Want to Be Free (among many others); he is the co-owner of Boing Boing, a special consultant to the Electronic Frontier Foundation, a visiting professor of Computer Science at the Open University and an MIT Media Lab Research Affiliate.
All opinions expressed by commentators are solely their own and do not reflect the opinions of Locus.
This article and more like it in the July 2020 issue of Locus.
While you are here, please take a moment to support Locus with a one-time or recurring donation. We rely on reader donations to keep the magazine and site going, and would like to keep the site paywall free, but WE NEED YOUR FINANCIAL SUPPORT to continue quality coverage of the science fiction and fantasy field.
Another pseudo intellectual of academia. Cory does not understand technology and what it can do in its most efficient configuration. Technology can form a Resource Based Economy that eliminates corporations, governments, and the concept of money.
The future is not so much about AI. It is about mechanical robotics that replaces human labor and frees us from the slavery of work and money. Real freedom is when the entire species is free of the bankers and the power structure of wealth. Automation can eliminate the ancient paradigm of the rich ruling over the poor. Something Cory is not seeing due to his limiting thinking.
The notion that the ‘efficient configuration’ of mechanical robots can somehow free us from the very concept of money, illustrates a depth of artificially intelligent understanding hitherto unseen online. Where can I subscribe to your newsletter?
There is absolutely no reason to think that the advent of robots will free us from the slavery of work and money. I mean, why would I give away the design of a perfectly good robot if I could charge you for it? Even if robots built robots, the design would still be proprietary.
So you want to get rid of proprietary rights to intellectual property? Good luck with that. These rights are being continually strengthened by the rich people who hold them now. An end of property can only happen through violent revolution, and it will be the poor fighting the robots of the rich, so ouch.
Look at car washes. Machines invented in the 70s now almost non existent -replaced by cheaper slave wage humans. Robots arent going to save the economy or humanity.
Interesting take on the situation, similar to how some nations surplus of people make it more advantageous to just pay someone to do a task instead of automating it and having to maintain the automation.
On the topic of a super-intelligent general AI’s role in potentially addressing these issues, would it not make sense to assume many of the manual labor jobs could be significantly reduced through technological advances if we manage to prop the world up long enough to get such an AI created? People would be needed as manual labor to build, say, a factory printing factory. After the initial bootstrapping, additional factories and their supporting infrastructure could be supported by a relatively small number of humans minding equipment.
Taking that out of the equation, the US had seen an issue with their citizens not wanting to take the available jobs left by migrant workers after a crackdown. Eventually necessity would force them into these less than ideal jobs, but a workforce used to retail or phone support jobs would likely not be suited for any manual labor jobs and not trained for the soft skills needed to treat trauma.
Not that I’m refuting any of these statements, I’m not qualified to do so, I just think it’s interesting to speculate on as we wait for the other shoe to drop.
Hello Tony Joy. You said, “the US had seen an issue with their citizens not wanting to take the available jobs left by migrant workers after a crackdown”. First, thank you for distinguishing between legal migrant workers who DO have seasonal visas to work in the United States versus so-called undocumented immigrants. Both are exploited, i.e. not paid living wages, enduring unsafe working conditions. Migrant workers aren’t subject to crackdowns though, because they are here legally. This arrangement has been in place at the request of US big businesses for decades.
US citizens would gladly do the work now done by seasonal migrants (and illegal aliens), if they were paid living wages. Until about the 1960s, Americans DID do such work, especially before consolidation of big business and agriculture. They raised families, built homes, etc. on living wages, while being protected by US workplace safety laws.
It is incumbent upon company management, whether private or public, to decide whether they care about their community’s and nation’s well-being more than excess profit taking. For example, even in New York City, my employer, S&P did not use a low-wage temp service for our company cafeteria. Instead, those workers were all full-time regular employees of S&P and received the same benefits, including health insurance and profit sharing, as the rating analysts, IT, and editorial staff. McGraw Hill owned S&P and decided to replace cafeteria employees with Sysco in 1999.
Next, I navigated to the link in your name and visited your website. I enjoyed reading several of your blog posts here! Your blog name’s subheading is wonderful, as it is sweetly self-aware and humble unlike most of the Internet:
Please consider adding the ability for readers to comment.
I see what you’re getting at, but I have several objections based on history.
First, it seems clear to me that the current popularity of “machine learning” is a nifty (and sometimes useful) parlor trick, but a detour from traditional AI research. We’ve hit dead ends before. Daniel Bobrow’s STUDENT was a neat idea which turned out not to be extensible to generalized language comprehension. It happens. But I don’t think that even today’s practitioners of ML ever thought that ML was a path to GAI. We always knew it was just a cul-de-sac. When it runs out, we’ll be back to what we were doing before. Only with must faster computers.
Second, it’s impossible to predict where and when breakthroughs will occur. Nobody saw a clerk in the Swiss patent office discovering general relativity. Nobody saw an operating system and its programming language, written by at AT&T programmer on a scavenged PDP-7 to play a computer game, taking over the computing world. Nobody saw a self-taught engineer with a molten candy bar in his pocket creating the microwave oven. These things always come out of left field. And during times of massive engineering efforts (like WWII, or the space race, or a forthcoming climate race), we have a million opportunities for discovery. A bull market is not prerequisite for discovery.
Third, if there’s one thing I’ve seen clearly this year, it’s that just because there’s a huge number of people out of work, and a huge number of jobs to be done, doesn’t mean those people will fill those jobs. They have to overcome issues with matchmaking and hiring and skills and training and certifications and geography and transportation and language and tools and supplies and everything else you can imagine. Yes, we’ll have 10 billion people and enough work for 10 billion workers, but that doesn’t mean we’ll have anywhere near full employment. No country has ever seen that kind of efficiency, and I see no reason this time will be any different. Even if a government was willing and able (ha!) to give everyone a job, there’s no way they’d be able to. Definitely not if there are any lobbyists left.
Great summary of the money supply and blind adherence. People put too much into AI due to movies and TV.
Pretty sure the US never ran out of ventilators in April-May, and certainly not California.
We didn’t run out of ventilators in April, May, or June in California or even in New York. I’m certain of that. Cuomo has so many extra ones in New York that he sent them to New Jersey and offered them to other states, who declined, as there was no need. Lack of ventilators isn’t the problem with coronavirus.
What is the purpose of labour (so called “lower” class)? – To produces goods for the owners / rich, either for their own consumption (caviar and Bentleys) or for sale at a profit (cheap goods for the lower class, medium goods for the middle class). In the absence of a need for human labour the need for a lower class does not exist.
What is the purpose of the managerial (middle) class? To manage the means of production and labour work force for the owners / rich. In the absence of the need for human labour the need for human mangers is nill.
In the end, with the dwindling resources of a finite planet the rich will not allow there own immiseration, and will thus seek to remove all goods and services beeing “squandered” on the unwashed masses. They will seek to exclude all members of the “lower” class from the functional economy. Those middle class workers who are purely managerial (unlike doctors, teachers, wet nurses – ones that provide a direct service to the owners / rich) will be added to the scrap heap of humanity.
Why would the owners / rich “share” their wealth with the poor when they offer no value?
Forget AI. Forget 30% unemployment, Forget relocating cities. The future is very much like the film “Elysium” (minus the space station).
6 billion people reduced to utter worthless poverty, living on the scrap heap that used to be the infrastructure of our world. And a few million people living in “gated” communities. Guarded by robotic soldier day and night, and pampered by the remnants of the middle class.
Thus the rise of “bullshit” jobs. This the fear porn of Covid-19. Thus the draconian lock down, and suppression of dissent. Thus the forty plus years of constant austerity politics.
I have a thought, an optimistic one. We have a pandemic of an incurable, untreatable disease, and nothing to protect us from it except quarantining, social distancing, hand washing and masks. Then we have a rate of lack of employment, housing, medical care and food, all matched by nothing since the Great Depression of 90 years ago. People can all agree that the two biggest wants are health and income. The questions are: can anyone have both, can everyone have both, and how do we go about insuring both to anyone and everyone?
Today’s technology is not helping us much, or at least it seems like that. Masks have been hard to get in the USA. Many people had to wait months to buy some. Automated robot labor can produce masks, gloves and other protective gear, and the US Postal Service can deliver this gear to the doorstep of every one of us. Do we have a shortage of helpful technology, one to neatly match our shortages of working people, medicines and testing kits, supplies, materials and money?
Technology can help many, but only of someone makes the decision to use that technology to help, and use it solely for monetary profit. Actually, we don’t need technology as much as we need people put to work. How great an untapped resource of able-bodied, able-minded people do we have in the USA? With all our closed businesses and unavailable goods and services, we would want enough people to get busy and “re-open” things. But none of us would want those helpful people to be exposed to the Coronavirus. All our problems can be solved and our needs met if people went to work on them. All except Covid-19 right now.
If there is a way to give jobs to all of the jobless and put every capable person to work, and do so while finding a way around the pandemic contagion so every worker is kept alive and well, then we have the possibility of living well with hope during the pandemic, and stay well when it ends. Solve the puzzle, and the miracle might happen.
”Keynes once proposed that we could jump-start an economy by paying half the unemployed people to dig holes and the other half to fill them in.”
It is too late even for Keynes. Irreversible climate change is here. Anyone with half-a-brain knows that. Trump & all neoliberals will deny this as ‘progress’ is still a live Capitalist operating mantra. Real progress is rejecting the 19th century modes of economic determinism and the ‘invisible hand’ that creates unsustainable levels of income and wealth accumulation which is directly responsible for the irreversible climate change as all resource accumulation denies sustainability. If the elite are allowed to maintain the status quo then the world is doomed to a dark age. Systemic change is necessary for the survival of humanity in a form that we can still call human and for the survival of the planet. It is down to us now. AI is utterly irrelevant.
Pretty spot on, in general, except for the climate panicmongering.
General AI is a myth. Automation replacing humans in any mass capacity is also a myth.
Even should one or both occur – past instances of real productivity improvements were quickly absorbed by everyone having higher expectations.
I must say that my past view of you was clearly wrong. Your work and commentary from a decade ago seemed technotopian to me – clearly not the case now.
But please put on the fireproof underwear as all the people who know nothing but believe in AI will attack you relentlessly.