Could Artificial Intelligence Finally Deliver the Promise of Central Planning?
Even the most sophisticated algorithms can only work with information that already exists—and meaningful economic information emerges from markets, not planning boards.
In the early 1970s, Chile embarked on one of the most ambitious economic experiments of the modern era.
Seeking to reshape the nation’s economy through technology, the country’s socialist administration enlisted British cybernetics expert Stafford Beer to create a computerized system capable of overseeing economic activity on a national scale. The initiative became known as Cybersyn.
Under the plan, factories throughout Chile would continuously transmit production data to a centralized network. Government officials, seated in a futuristic operations center lined with screens and control panels, would monitor economic performance in real time. Armed with enough information and computing capacity, planners believed they could guide the economy with scientific precision.
The vision embodied a long-cherished aspiration of socialist thinkers: the belief that advanced planning could outperform the spontaneous order of markets.
Yet Cybersyn never achieved the breakthrough its architects envisioned.
As the decade unfolded, the Chilean government imposed price controls on thousands of products while extending state ownership across major industries. The consequences were swift. Shortages became increasingly common, underground markets flourished, and the economy grew harder—not easier—to coordinate. Rising political tensions eventually culminated in the military coup of 1973.
To many observers, the conclusion appeared straightforward: centralized planning was incapable of reproducing the intricate coordination that markets accomplish every day.
Still, the dream refused to fade entirely.
The New Argument for AI-Driven Planning
The rapid rise of artificial intelligence has breathed new life into an old debate. If socialist planners of the past failed because their technology was too primitive, could modern AI finally overcome the obstacles that defeated them?
A number of contemporary commentators have suggested exactly that. In 2019, Jacobin published an article titled “Yes, a Planned Economy Can Actually Work,” arguing that vast datasets combined with powerful algorithms could potentially solve the classic calculation problems associated with socialism.
Even some economists have entertained the possibility. Before being awarded the Nobel Prize, Daron Acemoglu noted that advances in artificial intelligence might make centralized planning more viable than previously believed, implying that corruption could prove a greater challenge than technical feasibility.
At first glance, the argument carries weight. Artificial intelligence can absorb and analyze staggering amounts of information at speeds no human planner could ever match. Compared with the tools available to previous generations, today’s computational power is extraordinary.
But does greater computing power actually solve the fundamental problem?
To answer that question, it is necessary to revisit one of the defining economic debates of the twentieth century.
In 1920, Austrian economist Ludwig von Mises published his influential essay “Economic Calculation in the Socialist Commonwealth.”
Unlike many critics of socialism, Mises did not build his case around corruption, incompetence, or selfish incentives. Instead, he granted socialists their most favorable assumptions. Imagine, he proposed, that planners are highly intelligent, entirely benevolent, and sincerely devoted to public welfare.
Even then, he argued, the system would fail.
The reason centers on the indispensable role that private property and markets play in generating prices.
When the state owns the means of production, markets for capital goods cease to exist. Without markets, there can be no genuine prices for machinery, raw materials, equipment, and countless other productive resources. And without prices, rational economic calculation becomes impossible.
Prices are far more than numbers attached to goods. They are signals created through voluntary exchange, conveying information about scarcity, demand, and competing uses for resources.
Consider a straightforward construction decision: whether to build a floor using wood, ceramic, or marble. Prices instantly reveal valuable information about the availability of each material. If ceramic becomes scarce because demand rises elsewhere, its higher price encourages builders to consider alternatives. If timber supplies tighten, the rising cost of wood alters the calculation once again.
No one involved needs to understand every underlying cause. The price itself carries the essential message.
Remove that signal, and decision-making begins to resemble navigating through fog without a compass.
Several decades later, Friedrich Hayek expanded this critique in his landmark 1945 essay, “The Use of Knowledge in Society.”
Hayek argued that the central economic challenge is not one of computation but of knowledge. The information required to coordinate a modern economy is not stored in a single repository waiting to be processed. Instead, it is scattered among millions of individuals.
Much of this knowledge is deeply local. It reflects particular circumstances, shifting consumer preferences, temporary opportunities, specialized expertise, and practical experience rooted in specific times and places.
A significant portion is also tacit. People frequently possess useful knowledge they cannot fully articulate or quantify.
Markets solve this problem by continuously generating and transmitting information through prices. Central planning, by contrast, lacks an equivalent mechanism.
Some advocates of technology-driven planning assume Hayek’s argument merely reflected the computational limitations of the mid-twentieth century. But his critique was never primarily about processing power.
It was about institutions.
The crucial information needed for economic coordination does not exist in a ready-made database waiting for collection. Much of it comes into existence only through decentralized decision-making conducted within a framework of private property and voluntary exchange.
Artificial intelligence excels at analyzing data that already exists. What it cannot do is replace the decentralized processes that create that data in the first place.
An economy is not a machine that can be steered from a control room. It is a living, evolving system shaped by countless independent choices made every day by individuals, businesses, and consumers.
For that reason, artificial intelligence does not eliminate the socialist calculation problem. While algorithms may process information at unprecedented speed, they remain dependent on meaningful economic signals generated elsewhere.
For more than a century, every major technological breakthrough—from early computers to big data and now AI—has inspired predictions that governments would finally be able to manage complex economies from the center.
The technology changes. The argument remains remarkably familiar.
Artificial intelligence can undoubtedly analyze information with extraordinary efficiency.
But only markets possess the ability to discover that information in the first place.