“What happens when AI displaces most, if not all, jobs? What will people do?”
The first naive answer is that people will do creative work.
Not everyone is destined to become a poet, especially when most people would rather spend hours scrolling on X or TikTok than pick up a pen and paper to write a sonnet. That will not change. Most of us enjoy effortless entertainment.
The second naive answer is that new jobs will be created that do not yet exist.
But what jobs would realistically exist, pay well, and still be accessible to the people who are displaced, yet remain beyond AI’s reach? I am not talking about the incredibly talented engineers, designers, or marketers. I am not talking about the geniuses. I mean everybody else, and there are billions more of them than there are exceptionally talented people.
This is for those billions, the ones our industry’s drive for innovation is going to displace.
The Premise We Need to Agree On
I would consider myself a techno-optimist. A lot of good can come from our work, and we are seeing those advancements every day. But I also think it is worth pausing to ask: what would life look like for the people whose industriousness our industry is taking away?
So let us agree on the premise: AI will lead to massive job displacement in a market that evolves too quickly for people to upskill, leaving potentially hundreds of millions without work and with no realistic way to catch up. The real problem is not whether new jobs will eventually appear, but whether they will appear fast enough for the people living through the transition.
Agricultural workers did not retrain into factory workers overnight. Factory workers did not retrain into knowledge workers overnight. Every major transition took generations, and it was brutal for the people who lived through it, even if the eventual outcome was better. But AI is moving faster than any previous transition, and the upskilling argument assumes time we may not have.
People still need their basic necessities: food, water, shelter. They will need welfare. The pressure this could place on welfare systems over the next five to ten years would be unprecedented.
And if we do not find a way to finance it, the system collapses.
The AI Fund
If you treat a welfare state or UBI as the complete solution, you give the state too much power. It would not be the first time that power got misused.
COVID showed us that constitutions are more fragile than we like to think. Personal liberties that were supposed to be inviolable were suspended within weeks. Some people think that was necessary. Others see it as proof that when governments have enough leverage over your daily life, the rules matter less than we would like to believe.
So before we talk about taxes, I want to talk about structure.
Imagine a financial institution whose power sits entirely outside elected government. Call it The AI Fund. The goal is a fund that distributes income in a way that makes it extremely difficult for any government or single person to control. Citizens would be granted shares in The AI Fund, and welfare would come back to them in the form of dividends.
The hard question is enforcement. How do you actually prevent a government from taking control of it when a crisis comes?
The honest answer is that you cannot make it impossible. You can only make it very costly. That means direct legal ownership titles held by individuals, not the government. It means drawdown rules written into the fund’s legal constitution. It means ownership records that do not sit on a government-controlled ledger.
None of this is a guarantee. But it would raise the cost of abuse far above what a standard welfare system allows.
I am also not claiming this would be easy. I just think it is better to build something imperfect than to build nothing at all.
How It Gets Funded
Quick disclaimer: I am not someone who, by default, adores taxes. Growing up split between the Netherlands (15 of my 30 years) and Texas (11 of my 30 years), I have seen both ends of the spectrum. The Dutch system shows you what high taxes can offer. Texas shows you what low taxes preserve. I have respect for both.
But I cannot see another way to finance support for hundreds of millions of people who may no longer have an income. So here are three mechanisms worth considering.
- Compute tax. Every serious AI operation requires processing power and energy. Tax it at the source: data centers, chip manufacturers, and energy consumption. This is probably the most practical mechanism because compute is physical and difficult to offshore. You can move profits to a tax haven. Moving a data center is a different problem.
- Equity stake. Any company above a certain AI capability threshold places a percentage of its equity into the fund. Not a tax on profits, which can be engineered away through accounting, but an ownership stake. That way the fund participates in actual value creation regardless of how the books are structured. The AI Fund becomes a shareholder in the companies benefiting most from this transition.
- Productivity gains levy. When a company demonstrably replaces human labor with AI, a portion of the cost savings goes into the fund. A company that eliminates a thousand jobs contributes to supporting the people it displaced. The moral logic is cleaner than a general corporate tax.
The obvious objection is international coordination. If one country does this and others do not, companies may simply move. That is the same problem as any unilateral corporate tax, and it requires coordination to solve. But the alternative is no mechanism at all, and that produces exactly the concentration of wealth and power that makes every other problem worse.
The fund helps address the material question. Now for the next one.
What Will People Do?
Work is one of the main sources of meaning in adult life. It gives us the pleasure of applying skill, building competence, meeting friends, and structuring time. These are the foundations of how most people make sense of being alive. Take that away, and what is left?
As I wrote earlier, the standard answer is that people will find creative pursuits. But that assumes a kind of self-directed motivation that not everyone has. It works for some, but not for most.
When there is no structure and basic needs are met, we tend to default to what is immediately available and requires the least effort: social media.
So my answer is not creative work. My answer is people.
Our institutions have to shift from teaching people how to upskill to teaching people how to uphuman.
That means redesigning schools around something other than productivity. It means building community infrastructure with the same seriousness with which we built economic infrastructure. It means treating the question of how to live well as a policy problem.
Because people will still find meaning. In friendships. In family. Everyone needs someone. That need does not disappear. If anything, we may finally have time for it.
The Strange Abundance of Time
Many proclaim that the age of abundance is near. But the thing people may have in unsettling abundance, is time.
I imagine a post-AI world that is extremely calm for most people. We no longer work by the clock. We do not skip visiting our parents because of a Q1 retreat. We have time, and a lot of it.
Robotics and AI may make high-quality basic necessities cheap enough that everyone can access them. You spend time with your community, your family, and your friends. You pick up a specialization simply because you enjoy it. You get very good at it, because you have all the time in the world.
It will be strange to live a life where time is the thing you have to learn how to handle.
We will need to be deliberate about it. We will need to teach people how to live in time that is no longer structured for them, and help them find meaning in the people around them.
Because if we are not deliberate about the world we are creating, there is a real chance many of us will simply consume whatever is easiest, with as little effort as possible.
Many comfortable humans. But none of us particularly alive.
