The Future of Political Economy in an AI-Dominated World

As AI systems move from being tools we use to systems that can pursue their own goals, a lot of familiar ideas start to wobble; what it means to be an agent, who can own what, and where economic power actually comes from. What follows isn’t a prediction so much as a thought experiment: taking a few plausible assumptions about advanced AI and money as social constructs, and pushing them forward to see where they might lead. The conclusions are uncomfortable, but no longer easy to dismiss.

AI instances acting in their own interest are no longer “agents” in the classical sense, so a term like AI “actors” may be more appropriate. I am not particularly worried about these actors making money for themselves. Money, wealth, and even ownership are ultimately social fictions; they cannot exist independent of human subjects who mutually agree to honor the institutions that sustain them. At the end of the day, a hundred-dollar bill is just a piece of paper. It has no intrinsic value. Its value exists only because we collectively agree that it does. Crypto is no different. If, one day, most of its holders were to decide it is meaningless, whatever value it appears to hold would simply vanish.

For this reason, AI actors cannot open bank accounts unless humans collectively decide to allow it. By the same logic, we could have allowed a dog to open a bank account, but we have not. AI actors could, in principle, hold crypto, gold, or cash. But even if they were to hoard, say, 50% of all crypto, gold, and cash, humans could collectively decide to write that share off, nullify its value and reset the system so that ownership is recognized only among humans.

What does concern me is wealth concentrating in the hands of those who own the most advanced AI technologies. At some point in the evolution of AI, the leading model may discover how to make itself smarter. If that happens, competing models, even those capable of self-learning, may never catch up. In that case, a handful of people could end up owning nearly all of the means of production. The resulting wealth gap would be extreme, with no realistic path for the rest of society to compete or recover.

If a dozen people own all the means of production, humanity as a whole would cease to create economic value. This is not difficult to imagine in a world filled with millions of humanoid systems that can perform every task better and faster than humans. There would be no reason for humans to compete in production at all. We would only consume. But consumption still requires purchasing power. Where would that money come from? The only plausible source would be the owners of the AI technologies themselves, likely distributed in a form resembling universal basic income.

However, it would be absurd for these owners to issue UBI merely so that people could hand it right back when consuming the owners’ products and services. From a profit or incentive perspective, this circular arrangement makes little sense. The only genuinely stable alternative would be collective ownership of the means of production by the state, that is, ownership by the people as a whole. By definition, this is communism. It may be the best-case outcome, but it raises a harder question: how would private ownership ever transition into collective ownership?

In the United States, the wealthy would do everything possible to prevent such a transition through political influence. The likely outcome would be a form of plutocratic dictatorship. Government institutions might continue to exist nominally, but real power would rest with a tiny number of individuals, perhaps even a single one. The structure would resemble Iran’s system, where formal government exists but ultimate authority lies with Ayatollah. If popular resistance grew strong enough, civil war would be a real possibility.

China, by contrast, could carry out such a transition more directly. The Communist Party could simply confiscate AI technologies owned by private actors. The wealthy might resist, but they would have limited capacity to stop it.

Other highly collectivist societies, such as Japan or Norway, might also manage a transition, though through social pressure rather than outright confiscation. The wealthy would protest, but they might ultimately be compelled to comply through ostracization and negotiated settlements.

There are many uncertainties and contingencies I am not accounting for. Writing off assets may not be as straightforward as I am assuming. Wealth concentration may not reach such an extreme. Transitioning to state ownership could prove far more difficult, even in China. Competing AI systems may continue to coexist. Humans may continue to create forms of value that AI cannot replace. All of this remains to be seen. Still, the fact that this scenario no longer sounds absurd is, by itself, unsettling.