How will we be fed? That’s the biggest question not seriously being addressed amid all this talk about whether or not artificial intelligence will end up taking over all of our jobs.
Formidable though the technology appears, similar fears have popped up repeatedly since the Industrial Revolution, and most working-age adults remain employed. Still, what is sorely missing is a serious debate about what to do if this future in fact materializes.
For Open AI’s Sam Altman “the future can be vastly better than the present” because AI will make us stinking rich. But that seems like a risky assumption, for almost everyone except Altman and his fellow techno-oligarchs.
Even if AI generates enormous economic prosperity, its distribution will remain a political challenge. This juncture calls for a serious, open debate about how the fruits of this prosperity will be apportioned among humanity.
Addressing the question has two parts. The first is about how to design a technically efficacious system to redistribute the fruits of the economy as machines take over and labor’s share of income drops eventually near zero.
The more important question, though, is about how this economic reorganization will restructure power. Who will decide what to tax once AI destroys labor income, which provides the main source of government revenue in most advanced countries? Who decides how much everyday people who do not have an equity stake in the AI revolution get to consume?
How will society be organized in a world in which machines generate most or all economic output and a few dozen techno-billionaires get to decide what share of the world’s resources – money, energy, minerals – should be allocated to further expand superhuman intelligence? Who else gets a say on whether to direct more resources to, say, healthcare or agriculture, or education instead?
“We need guardrails that preserve human agency, human oversight and human accountability,” noted United Nations secretary general, António Guterres, at the AI Impact Summit in New Delhi last week. The future of AI “cannot be decided by a handful of countries or left to the whims of a few billionaires”.
In AI circles, there is a lively debate about the “alignment” challenge – ensuring that the machines operate in ways that serve the goals of whoever runs them. The bigger challenge is to align the goals of AI systems and their owners with the broader goals of society. AIs will do lots of things of consequence to all of us. Our democratic governance tools seem too feeble to constrain the urges of the oligarchs at the helm of these new technologies.
Technological change drove democracy’s spread around the world as the rise of an urban working class proved indispensable to the economy and political systems adjusted to represent them. But if the work of ordinary people becomes irrelevant, what happens to people’s power to affect their system of government?
Anton Korinek and Lee Lockwood of the University of Virginia put together a primer with ideas about how public finance might work in the AI era. They propose that consumer taxes will pick up the slack at first, as labor income shrinks to zero. In a world dominated by artificial superintelligence, though, the footprint of human consumption would shrink as most of the returns from machines’ economic output were reinvested, calling for a tax on capital to shoulder most of the burden.
Maybe taxes could also be used to slow down the transition. Another idea discussed by Korinek and Joe Stiglitz from Columbia University is that in early stages, when human labor retains its relevance, taxes could be used to steer tech investments toward technologies that help workers do their jobs better rather than replace them. Korinek and Lockwood propose other taxes, on fixed factors like land, spectrum or data, or monopoly rents, that add nothing to society’s wellbeing.
It sounds doable. The problem is that the owners of these disruptive technologies must be convinced to do something that does not come naturally to them: share. Taxes in the US amount to less than 26% of GDP, eight percentage points less than the OECD average. Capital taxation amounts to just over 2% of GDP. These numbers will have to go much higher, since people will no longer have wages to live on and will rely more heavily on government largesse.
Don’t hold your breath. The OECD’s global tax deal, finalized in 2021, was designed to curtail the ability of American tech companies such as Amazon, Google and Meta to engage in tax shifting, parking profits in the lowest tax jurisdiction they could find. But while the Biden administration was broadly supportive of the deal, Donald Trump – whose campaign benefited from nearly $400m in donations from various tech oligarchs – unilaterally withdrew in early 2025.
Unusual ideas may be called for to keep society afloat, given the scale and breadth of the expected AI revolution. One would be to directly distribute the equity of artificial intelligence ventures. Taxes might be collected in shares rather than cash, to amass a public stake over time. Rather than tax the returns on AI investments, a more radical proposal would be for the government to expropriate a chunk of equity upfront to redistribute among the population and directly grant Americans a share in AI’s promised cornucopia.
“If AI development stalls, returns remain modest; if AI transforms the economy, returns are likely to rise,” Korineck and Lockwood wrote. “This automatic adjustment proves valuable given radical uncertainty surrounding AI development.”
But these big ideas face big challenges. Governments will have to act before artificial intelligence gets too big, which seems unlikely in the current climate.
The technology oligarchs at the helm of this revolution have also vigorously resisted government efforts to curb their power or take their cash. Despite her best efforts, Silicon Valley nemesis Lina Kahn was largely unable to crack tech monopolies during her stint as President Joe Biden’s main trustbuster at the head of the Federal Trade Commission.
In the meantime, the moneyed in Silicon Valley are not only mobilizing vast resources to steer American politics. As a plan B they are working to build their own “network-states”, whether in Greenland or Nigeria, Honduras or the Caribbean island of Nevis, hoping to evade democratic governance if they can’t get their way under America’s democracy.
Who knows what they might do when they have replaced all human labor. If artificial intelligence grows as powerful as Silicon Valley’s oligarchs expect it to become, the only available strategy to keep us all fed in the world after work might be to go hat in hand and ask the moguls, politely.