Why AI Won’t Leave Workers Starving Amid Robot-Made Abundance

A popular fear about artificial intelligence goes something like this: machines will become so productive that they generate staggering abundance, while simultaneously throwing most people out of work and leaving them unable to afford anything the machines produce. Call it the AI economic apocalypse theory.

The idea holds that AI will first absorb cognitive work, then move on to farming, manufacturing, and services, performing nearly everything humans do today with far greater efficiency. The owners of these systems become immensely wealthy. Everyone else, having lost their jobs to automation, supposedly loses their purchasing power along with it.

This prediction rests on two claims that don’t fit together. It says people will be shut out of the bounty that automated production creates, yet it also implies that same production destroys people’s ability to trade with one another. In other words, we’re told to expect a world of AI-driven abundance that is somehow also a world where most people go without.

This isn’t a fringe idea dismissed by serious people — it’s a scenario floated by plenty of thinkers, including some who work inside AI companies themselves.

What’s missing from the theory is any explanation for why people with unmet needs, working skills, and productive resources would simply stop trading with each other. Why would anyone sit idle while neighbors and fellow citizens still need the goods and services people are fully capable of providing one another?

The Contradiction at the Core

The apocalypse theory can’t have it both ways. If AI performs all labor and delivers goods and services at dramatically lower cost, that isn’t the deprivation scenario doomsayers warn about — it’s abundance. The very mechanism that supposedly destroys the market for human labor is the same mechanism that makes life cheaper for everyone.

Consider the alternate version of the fear: that AI’s owners hoard the output or price it out of reach for ordinary households, retreating into a kind of walled-off prosperity while everyone else is locked out. Even in that scenario, people still need food, shelter, clothing, transportation, education, and entertainment. They still have their knowledge, their skills, their tools, and their ability to organize and trade. If AI-driven abundance is withheld from them, they’re simply back in an economy that functions the way economies always have — through people supplying one another.

Part of the confusion comes from mistaking companies for the actual source of jobs and income. Companies are middlemen — they coordinate supply and demand for labor and goods. They don’t create the underlying human need for exchange. If AI-owning firms fire everyone and sit on their output, other entrepreneurs will simply organize new ways to connect people who have something to offer with people who need it.

Cheap Goods Mean More Purchasing Power

Now consider the scenario where automated businesses do make their output broadly available, at prices that undercut human producers. Specific jobs disappear. Incomes for some may fall. But the goods and services those jobs used to provide become dramatically cheaper as a result — and that drop in prices has to be weighed against the loss of income, not ignored.

The apocalypse argument wants it both ways here too: it uses cheap automated output to explain why human workers can’t compete, then quietly drops that same cheap output when it comes time to assess what ordinary households can actually afford. That isn’t just unlikely — it’s internally inconsistent.

We Are Not Horses

Defenders of the doom scenario often point to horses, whose labor market was wiped out by the combustion engine and mechanized farming. The argument goes that humans could meet the same fate once machines can do everything we do.

But horses had a problem people don’t have: they couldn’t trade with each other. When human demand for horse labor collapsed, horses had no way to build their own economy among themselves. If AI makes human labor superfluous to AI companies, people aren’t left as helpless as horses were — they retain the ability to supply each other with goods and services, and that ability generates its own means of payment. The horse economy existed to serve human needs; when horses stopped being useful for that, they were finished. The human economy exists to serve human needs, full stop, and humans can keep serving those needs for each other even if AI doesn’t need them at all.

What Abundance Would Actually Look Like

Taken to its logical extreme, this abundance scenario means basic necessities could become so cheap that people need far less income to live well. A falling dollar wage doesn’t necessarily mean a falling standard of living — what matters is what that wage can buy.

It’s true that in such a world, the owners of automated capital could grow enormously richer while everyone else also enjoys historically unmatched material comfort. That may widen measured inequality, but it’s hard to see why that alone should trouble anyone, since everyone would be better off in absolute terms.

A total abundance scenario, where every want is satisfied at virtually no cost, is unlikely. Wherever goods and services remain scarce or expensive, people will still have reason to supply them; wherever automation makes things cheap, households will simply redirect their spending elsewhere. Craftsmanship, personal service, entertainment, and human attention are likely to hold or even gain value, and new wants will emerge as old needs stop consuming so much of people’s budgets.

In fact, human labor could become more valuable, not less, in a world of cheap necessities. If people can meet their basic needs cheaply, many may choose to work less, raising the