What Happens When Human Labor Becomes Optional?
Most conversations about AI and work start with the same question: will it take my job?
That’s a reasonable thing to worry about. But if we imagine genuinely capable artificial general intelligence (AGI), I think that question only gets us part of the way there.
I’m talking about systems that can learn, reason, plan, use computers, conduct research, design products, and manage businesses. Eventually, perhaps, they could also operate robots capable of doing most physical work.
If that happens, human labor may stop being the main constraint on what we can produce. We could reach a point where society no longer needs most people to work in order to meet its material needs.
Plenty of people would still want to work. The difference is whether the economy needs them to.
From doing more to needing fewer people
A rough picture of today’s economy looks like this:
Resources + capital + human labor → goods and services
An economy built around advanced AI and robotics could look more like this:
Resources + energy + machines + AI → goods and services
These are simplifications, of course. But they help explain why there’s a big difference between AI making us more productive and AI making our labor unnecessary.
I can imagine a few ways this could play out.
First, AI could remain mostly a tool that makes people dramatically more productive. A software engineer might direct hundreds of AI agents. A scientist could run thousands of virtual experiments. One entrepreneur might operate a business that currently needs 500 employees.
People would still work, but far fewer of them might be needed to produce the same output.
A further step would be the automation of most intellectual labor. Programming, accounting, law, design, administration, analysis, marketing, customer service, research, and management could become activities largely handled by AI. Humans might still own the businesses and make high-level decisions, while contributing very little of the labor needed to run them.
Then there’s robotics. This is the threshold that would change things most.
If affordable, capable robots can build houses, farm, manufacture goods, repair infrastructure, move supplies, cook, clean, mine resources, and maintain other machines, much of the production cycle could run without us. Human labor might genuinely become optional for meeting society’s material needs.
We’ve automated jobs before. Usually, though, people could move into other kinds of work. Humans were still the adaptable, general-purpose workers the economy relied on.
AGI combined with general-purpose robotics could automate that role too. At that point, I’d be asking why a human has to do a job at all, rather than which job they should move into next.
People would still have things worth doing
There’s an easy leap from “the economy doesn’t need your labor” to “you aren’t needed.” I don’t think that follows.
We value plenty of activities because another person is doing them. Someone might prefer a human therapist, teacher, musician, chef, doctor, or caregiver even if a machine performs better at many parts of the job. The same could be true of artisans, athletes, priests, and politicians. Ability isn’t always the only thing we care about.
Handmade goods are an everyday example. A machine can produce something more consistent, more quickly, and at a lower cost. People still pay extra for an object someone made by hand. Its imperfections can be part of the appeal.
Chess is another useful example. Computers have been better than humans for a long time, and human chess is still interesting. We watch people play because we care about the players, their decisions, and what they’re capable of.
Imagine a robot that could run 100 metres in three seconds. That wouldn’t make the Olympics pointless. We want to see what a human can do.
Painting would still be worth doing if AI could generate a masterpiece. Raising your child would still matter if an AI nanny could theoretically do a better job. Building a company might become something people do for the same reason they climb mountains: they want to find out what they can accomplish.
More of our activity could be driven by curiosity, relationships, competition, pride, and self-expression. We’d have more room to do things because we want to, rather than because we need the income.
Who owns the machines?
For me, this is the hardest economic question in the whole picture.
Suppose an automated factory can produce almost everything people need. AI operates it, robots do the physical work, and very few employees are required.
Who owns that factory?
If ownership is concentrated in a few hands, we could end up with enormous productive capacity and a population that has very little bargaining power.
Today, labor gives ordinary people something to bargain with. You need my time and skills, so you have to pay me. That arrangement is far from fair or reliable for everyone, but it’s how many people get a share of what the economy produces.
If machines no longer need our labor, that source of leverage weakens. A factory’s ability to make more things doesn’t, by itself, give everyone the means to buy them.
We’d need some other way to distribute purchasing power or provide access to what people need. That could mean universal basic income, universal basic services, broader capital ownership, public ownership of some AI infrastructure, sovereign wealth funds, or dividends from automated production. Falling prices could help too. The eventual arrangement might be something we haven’t thought of yet.
The important point is that two societies with the same technology could end up looking very different.
In one, housing, food, healthcare, education, transport, entertainment, and energy could become affordable enough that people work ten hours a week, or don’t need paid work at all. They could spend more time raising families, learning, doing science, playing sports, creating things, exploring, and building communities.
In another, the technology could generate tremendous wealth while most of the benefits go to a small group of owners.
Having the technology wouldn’t settle the question of who gets to benefit from it.
Some things would still be scarce
Even extraordinary automation wouldn’t make everything abundant.
Land in desirable locations would still be limited. So would access to natural environments, unique physical objects, tickets to a live event, and time with particular people. Energy and computing capacity might remain constraints too.
You could build ten billion excellent houses. You couldn’t manufacture another beachfront Vancouver.
Human attention, social status, and political power would still be contested. If everyone had a comfortable home and everything they needed, people might compete even harder over reputation, influence, relationships, achievements, and experiences.
Economics would still matter. Its focus could shift away from shortages of labor and manufactured goods toward the things we can’t easily reproduce.
A comfortable life wouldn’t necessarily be a life without competition.
What would we do with ourselves?
There’s also a personal question here that I think gets underestimated.
For much of human history, survival has imposed a structure on people’s lives. Hunt, farm, work, earn, survive. You might dislike the routine, but you usually don’t have the option to abandon it.
If earning a living were no longer necessary, more people would have to face a question that some wealthy people already encounter:
What do I actually want to do with my life?
Some would flourish. They’d study mathematics, compose music, raise children, travel, garden, volunteer, conduct research, build strange little projects, compete in sports, explore space, or spend ten years getting good at woodworking.
Others might struggle. Work provides more than a paycheck. For many people, it’s a source of structure, identity, social contact, and a feeling of being useful. Removing the need to work wouldn’t automatically give them something satisfying to put in its place.
An abundant world could solve a lot of material problems and still leave people feeling lost. We’d have to take that seriously too.
A bigger change than job automation
My rough way of thinking about this is:
- In the near term, AI changes how we work.
- More capable AI agents change how many people are needed to do the work.
- AGI and mature robotics could change whether human labor is economically necessary at all.
None of that is guaranteed. But if we reach the last stage, “AI took all the jobs” might eventually sound a bit like “tractors took away subsistence farming.”
There’s truth in the complaint. The disruption is real, and it matters enormously to the people living through it. But the lasting change was that civilization no longer required almost everyone to spend their lives producing food.
AGI could potentially do something similar for labor itself.
I think many people would still work. They’d still build businesses, make things, care for others, and try to do something difficult well.
What interests me most is whether living comfortably could stop depending on being economically useful to somebody else.
That would change far more than how we work. It would change how we live.