No, AI Won’t Make Money Obsolete

Authored by Peter Earle via the American Institute for Economic Research,

The notion that artificial intelligence (AI) at full bloom might eliminate the need for money reflects a deep confusion about what money is and does. Money is not merely a barter-avoiding convenience layered onto an otherwise frictionless world. It is a solution to fundamental problems of exchange, profound difficulties in coordination, and comparison of alternatives under scarcity. Even in a hypothetical future defined by extraordinary productivity gains and broadly collapsing prices, those underlying problems do not disappear. Instead, they change form, and for as long as scarcity, tradeoffs, and uncertainty persist in any domain, so too will the need for money.

To begin with, the most basic point: Scarcity is not abolished by abundance. It is displaced. AI may dramatically reduce the cost of producing many goods and services, particularly those that are digital or easily replicable. But large swaths of economic life remain governed by constraints vastly beyond the power of computation. Land is fixed. Location is inherently scarce. Prime real estate in places like New York City or Tokyo will not become abundant simply because construction costs fall precipitously. The same holds for proximity to infrastructure, culture, or social networks. These are rival, excludable goods, and in such conditions, exchange requires a mechanism for allocating access. Money remains the most efficient one yet developed.

Time is another irreducible constraint. Human attention, especially in its highest-value forms, cannot and will not scale infinitely. The time of a skilled surgeon, an experienced trial lawyer, or a sought-after performer remains finite, tentative, and rivalrous. Even if AI were to augment a person’s capabilities, it does not eliminate the fact that the person’s attention must be allocated among competing uses. The same applies to live experiences: concerts, events, one-on-one advisory relationships, and so on, where presence itself is scarce. In such contexts, prices are not a relic—they are a reflection of incontrovertible limitations.

In fact, abundance often amplifies the importance of scarcity. As mass-produced goods become ever cheaper, a premium will shift toward what cannot be easily replicated. Consider status goods, fixed positional assets, and signals of taste. Luxury brands, rare collectibles, and authenticated works derive value precisely from their limited supply and provenance. If AI floods the world with high-quality substitutes, the value of the original or source item may increase, not decrease. Money, in this sense, becomes a way of expressing relative preference over increasingly differentiated manifestations of scarcity.

Physical systems themselves impose limits. Energy, for example, may become cheaper on average, but it inexorably faces capacity constraints, especially during peak demand periods. The same is true of certain materials, bandwidth, and computational resources in periods of congestion. Even highly advanced systems must allocate finite capacity across competing uses, and money prices remain an extraordinarily efficient, indeed elegant way of doing so. Without them, the hallmarks of rationing—queues, quotas, and administrative fiat—appear, none of which eliminate, but rather contend with and obscure, scarcity.

The unavoidable force of uncertainty is perhaps the most decisive argument against the obsolescence of money. Risk does not vanish amid colossal gains in productivity and output; if anything, complex, tightly coupled systems generate new forms of it. An explosion of goods and services will tax resources, time, and human capital, which will in turn generate new forms of insurance, hedging, and credit to transfer and price risk. Those functions require not just a medium of exchange but a unit of account to operate effectively. The idea that AI could eliminate uncertainty is as implausible as the idea that it could eliminate time.

Institutional realities reinforce the former point. Anywhere one finds government or governance, excludability inevitably follows. Property rights, regulatory approvals, access to bespoke networks, and enforcement mechanisms all create domains in which access is controlled. Money is readily suited to become (or, in fact, continue to be) the means by which access is negotiated, transferred, or prioritized. In a world inundated by output, trust and verification become more valuable. Certification, auditing, and reputation systems all rely on mechanisms of exchange that presuppose some form of monetary unit.

Money also plays a central role in coordinating urgency and priority. When resources are scarce in time versus in quantity (faster service, guaranteed delivery, and dedicated capacity), money allows individuals to signal how much they value immediacy relative to others. Absent money, such decisions do not disappear; they are made through other, often less transparent means.

An AI-driven deflationary boom would likely compress the prices of many goods and services. Perhaps dramatically so. But that would not—and could not—eliminate the need for money. It would shift the domain in which prices and calculations operate toward the nonreplicable, capacity-limited, and institutionally governed.

Money does not disappear in the face of abundance; it instead follows scarcity wherever it emerges.