The Dangerous Passage to Abundance

AI may eventually make most human labor unnecessary, and the age of ownership may turn out to be a temporary stage. The period in between is the dangerous part. Even if technological abundance becomes possible, there is no reason to assume the transition to it will be peaceful. The machines capable of producing abundance will not appear already owned equally by humanity. They are being built inside the institutions that exist today: corporations, nation-states, financial markets, militaries, legal systems and political movements. Those institutions compete, defend power, protect property and distrust one another. The difficult period may therefore come after AI becomes extraordinarily capable but before society has adapted to what that capability means. That is the world through which a child born in 2026 will reach adulthood.

Technology can change faster than institutions

The pattern is familiar. Industrialization transformed production faster than labor law adapted to factories. Financial markets became global faster than regulation did. Nuclear weapons appeared before international institutions understood how to govern a world in which cities could be destroyed in minutes. The Internet transformed communication before societies understood what ubiquitous networks would do to privacy, media and politics. AI compresses several such transformations into one technology, affecting economic production, scientific research, military capability, cybersecurity, information, medicine, education and government.

The danger is not merely that AI becomes powerful. It is that capability grows faster than the institutions responsible for controlling and distributing it.

Jobs can disappear before the income system changes

The first conflict is economic. Today’s system assumes people obtain access to society’s output through income, and for most adults income comes from work. As AI raises productivity, a company that once needed 1,000 people needs 500, then 200, then 50. It produces more than before. Society becomes richer. Fewer people receive wages for creating that wealth.

At first this looks like ordinary technological change: workers retrain, new occupations appear, productivity creates new industries, as it has for centuries. AI creates a harder problem if it can also perform the new occupations into which displaced workers would normally move. Then society meets an unfamiliar contradiction: the economy can produce enough, but it no longer needs enough workers to distribute purchasing power through employment. Technology has solved the production problem before politics has solved the distribution problem.

So far the evidence is contested, but for the young it points one way. Stanford’s Digital Economy Lab, using payroll data from ADP covering millions of American workers, reported in August 2026 that employment of 22 to 25 year olds in the most AI-exposed occupations had fallen about 11% since late 2022 while that of their peers in the least-exposed occupations grew about 10%, leaving the exposed group 19% below trend. Experienced workers show no such gap, and the effect runs through reduced hiring rather than layoffs.² Yale’s Budget Lab, looking at the labor market as a whole, found no discernible disruption in the three years after ChatGPT’s release, and New York Fed economists attribute most of the rise in graduate unemployment to remote work rather than AI.³ Whichever reading is right, the mechanism described here has not arrived at scale. The question is what happens if it does.

Ownership therefore becomes power

The automated systems still have owners. As AI and robotics generate a growing share of national output, the companies that own those systems become extraordinarily valuable and their shareholders hold claims on a growing fraction of productive capacity. Nothing about this is illegitimate under today’s rules.

But its political meaning changes as labor matters less. The old bargain was roughly that capital needs labor, labor needs capital, and gains are distributed between them through wages, profits, taxes and political bargaining. When capital no longer needs labor, workers lose not just jobs but bargaining power. A society could become technologically capable of abundance while economically producing extraordinary concentration. The central dispute stops being about tax rates or wages. It becomes:

Who should own the productive intelligence?

Today’s political ideologies are built for the transition, not the endpoint

Here contemporary political arguments start to look historically narrow. The American right emphasizes property rights, markets, entrepreneurship and private capital. The mainstream left emphasizes worker protection, redistribution and public investment. Democratic socialists emphasize worker or social ownership of productive assets. During the next twenty years those differences may matter a great deal.

But all three assume an economy in which human labor remains central. The right says people should support themselves through productive participation in markets. The conventional left says workers should receive fairer compensation and stronger security. The socialist left says workers should control more of what their labor produces. If machines create most economic output, each premise weakens. “Workers should receive more of what they produce” means less if they produce little of the total. “Everyone should support themselves through work” is harder to defend if society does not require everyone’s work. And unrestricted private ownership becomes politically explosive if a few people control systems that constitute society’s entire productive base. The argument may reorganize around a more basic question: What claim does citizenship itself confer on automated abundance? No settled ideology answers it yet.

Power rarely dissolves because it has become unnecessary

There is a comforting assumption in many discussions of AI abundance: once the machines can produce enough for everyone, the owners will simply share. History gives little reason to expect this. Workers did not receive labor protections because industrial owners found them philosophically elegant. Monarchies did not disappear because representative government had become imaginable. Colonial systems did not dissolve because administrators concluded self-government was morally superior. Institutions change through pressure, negotiation, elections, organization, and sometimes violence.

Why should control over advanced AI be different? Those who own automated productive systems may regard their claims as perfectly legitimate, and under existing law they may be right. The conflict appears when ordinary property rights meet an extraordinary condition: a relatively small set of privately owned systems becomes capable of producing most of what civilization needs. Then a question that once concerned wealth begins to concern sovereignty.

Governments will compete before abundance makes competition less important

The same problem exists between states. AI is not being developed by a single global civilization but inside competing countries. The United States and China both want technological leadership. India wants capability and strategic autonomy. European states want competitiveness without surrendering political control. If advanced AI offers major economic or military advantage, no serious government can stop competing while it believes rivals will continue. Country A accelerates because it fears Country B; B observes and concludes it must move faster. Each may sincerely believe it is acting defensively, and together they create an arms race. This is already visible in cybersecurity, intelligence analysis and increasingly autonomous military systems.¹ The technology that might eventually make material resources less scarce can increase geopolitical competition on the way there.

Warfare may become easier to scale

AI also changes the economics of force. Modern war increasingly runs on drones, machine vision, electronic warfare, autonomous navigation and software-driven targeting. Ukraine, which built a few thousand first-person-view drones a year in 2022, produced about 3 million in 2025 and aims for more than 8 million in 2026.⁴ The Center for Strategic and International Studies estimates that Russia’s casualty ratio against Ukraine rose from between 2 and 3 to 1 for most of the war to nearly 8 to 1 in the first half of 2026, in large part because of Ukrainian drones, including AI-enabled ones.⁵ As these systems improve, one historical constraint on military power weakens: the need to risk large numbers of one’s own people. Autonomous systems do not remove that constraint, but they reduce it for the side deploying them. Cyber capability adds another asymmetry. OpenAI says Astra is the first of its models to reach what it calls the Critical level of cybersecurity capability, able to find previously unknown security flaws and develop exploits for them without a person guiding each step.⁶ A small organization with such a system might conduct operations that once required a large intelligence or military establishment.

This does not mean AI inevitably causes more war. Eventually, abundance could weaken some motives for conquest. But during the transition, AI increases the capability available to actors operating under today’s motives.

The weapons of the future arrive before the politics of abundance.

Religion and morality enter where engineering stops

The transition will not be fought only over money and military power. AI will increasingly shape medicine, education and biology, where the hard questions are not technical. AI may help develop better gene therapies and far more precise personalized medicine. Over time the boundary between treating disease and enhancing normal human capability becomes harder to draw, well within the lifetime of a child born today. Should a genetic intervention be allowed to prevent a serious inherited disease? What if the same technology can later improve an ordinary trait? Should wealthy families have early access to enhancements unavailable to others?

Science can answer whether we can do this. It cannot answer whether we should. Religious traditions, secular ethics, courts and political systems will answer differently, and those disagreements become politically consequential because AI expands what is technically possible faster than cultures converge on what is permissible.

Abundance will arrive unevenly

Scarcity will not disappear everywhere at once. Information, intelligence and software may become extraordinarily cheap; education and manufacturing may follow as robotics improves. But land in desirable places, political authority and certain raw materials will remain scarce. Energy infrastructure takes time to build. Healthcare may become technologically abundant before access becomes institutionally abundant. Some countries will automate faster than others, and some people will own far more AI-linked capital than others. The transition will therefore produce a strange mixture, abundance in capability alongside scarcity in access, which is exactly the kind of mismatch that produces political tension.

The paradox of the transition

AI could eventually weaken many traditional causes of conflict. If energy is abundant, fossil-fuel reserves matter less. If food production is automated, fertile land matters less. If manufacturing is cheap and distributed, distant industrial capacity matters less. If societies can generate extraordinary wealth domestically, conquest becomes less attractive.

But the route to that world may first intensify competition. States compete for the technology, companies for ownership, workers for economic security, political movements over distribution, cultures over acceptable uses. The technology capable of reducing scarcity first makes control of the technology itself the most important scarce resource.

The transition is the real political problem

A child born in 2026 turns eighteen in 2044 and thirty in 2056. She could plausibly grow up watching AI move from a tool used by workers to an infrastructure that performs much of the work itself, while the old system remains legally and politically intact and the new productive system emerges underneath it. Jobs may decline before income systems change. Automated wealth may grow before ownership broadens. Military capability may expand before arms-control institutions adapt. Biological technology may advance before societies agree on its limits.

The hardest question therefore lies not at the endpoint but in the passage: Can humanity move from one system to the other without allowing the transition to become more destructive than the abundance is valuable? There is no guarantee. That is why today’s political choices matter even if today’s political categories do not survive. The institutions built over the next few decades may determine who owns AI capital, how displaced labor retains economic security, how automated wealth is distributed, what governments may delegate to machines, which military applications are prohibited, and what biological interventions societies permit.

The child born today may ultimately inherit abundance. Before she gets there, her generation may have to answer a much older question:

Can those who acquire extraordinary power be persuaded to share it before others are forced to take it?

That, not the technology itself, may be the hardest part of the journey.

Sources