American University behavioral economist Alex Imas recently published a long-form article titled 《What will become scarce?(What will be scarce?)》, citing economic and psychology theories to emphasize that AI automation will not completely replace human jobs; instead, it will create a new economy centered on “emotion and relationships,” turning it into the scarce commodity after the value shift in the post-AI era.
Starbucks’ failed experiment reveals automation blind spots
The article opens with Starbucks. This coffeehouse chain giant, with a market value of more than $112 billion, in recent years has tried to squeeze costs by reducing employees and expanding automation—yet the result was the opposite of what it intended. CEO Brian Niccol publicly admitted this was a mistake: “The handwritten text on the cups, ceramic mugs, comfortable seating, and other details and hospitality—these are the key reasons customers are willing to stay awhile and keep coming back.”
In response, the company immediately decided to increase the number of baristas at each store and scale back the automation. Imas uses this as a lead-in and raises the core question:
If machines can do most of what most humans can do, what scarcity is left?
His answer is: scarcity still exists, but its types and positioning will undergo a fundamental transformation.
From agriculture to factories: history reveals how the economy “retools itself”
To understand the transformation in the AI era, you must first look back at the historical规律 of “structural change.” In 1900, about 40% of the U.S. workforce worked in agriculture; today, it’s less than 2%. Yet agriculture has not disappeared—it’s just that automation lowers costs and increases output, causing the share of spending on food to fall. The workforce then shifts to manufacturing, and later to the service sector.
1850 to 2050 structural employment transitions: agriculture (green), manufacturing (light blue), service sector (dark blue) each saw expansion and contraction; now it has come to the AGI era (deep red)
In economics, there’s a term called “Baumol’s cost disease (Baumol’s cost disease),” which refers to the idea that in sectors where productivity grows quickly, things become cheaper, but demand for these goods has a ceiling. By contrast, sectors where productivity is difficult to improve—such as live performances, personal care, and education—although relatively expensive, can continue to absorb more spending and employment.
As a 2021 study in the academic journal Econometrica found, once people become wealthier, they don’t just buy more of the same things—they also shift toward goods and services with higher income elasticity.
Imas believes AI is replaying this script, just on a larger scale and at a faster pace.
Imitation desire and scarcity: people will always pursue what others can’t get
Imas also cites French thinker René Girard’s theory of “Mimetic Desire (Mimetic Desire).” Girard argues that when people desire something, it’s often not because of its functional value, but because others also want it yet can’t obtain it. Status and the very feeling of exclusivity are the fuel for human desire.
Human-made vs. AI-made: a comparison of willingness-to-pay for artworks
In addition, Imas’ research finds that AI involvement significantly weakens the premium associated with exclusivity: human handmade art can command a 44% value premium due to scarcity, while AI-generated works, even if they claim to be limited, have a premium of only about 21%. The reason is that AI makes items feel “as if they can always be copied,” undermining the psychological foundation of scarcity.
This logic applies not only to art. Across education, healthcare, psychological counseling, hospitality services, and crafts—any domain in which “who provides it” is itself part of the experience will hold true as well. A person’s presence, judgment, and warmth are things that AI struggles to replace.
Where will future career paths be? “Emotion and relationships” practitioners become the biggest winners
Taken together, Imas outlines what he calls the “Relational Sector (Relational Sector),” an economic segment centered on emotional connection and exclusive experiential value. As AI drives down prices for standardized goods and real incomes rise, people will spend more money on jobs like nurses, therapists, teachers, artisans, and live performers—potentially even creating a wave of new jobs that don’t exist yet.
However, Imas also admits that this framework mainly applies to developed countries. For developing countries that have long relied on exporting commodities, the disruption brought by the AI wave will be more complex and severe.
Biggest controversy: the issue of income distribution remains unresolved
However, the article also sparked some questions on X. One reader pointed out: “This argument seems to rest on an unstable premise—that in the post-AGI era, the economy will still be led by humans.”
If AI comes to account for 90% or more of the overall economy, and human income shrinks drastically, then there won’t be enough purchasing power to support markets for craft beer or bespoke tailored suits.
This criticism also mirrors the biggest gap in Imas’ theory: Imas’ framework analyzes from the demand side, explaining what people want after their incomes rise, but it skips over the allocation side of the problem.
In a scenario where AI accelerates the concentration of capital, whether the post-AGI era will include supporting measures such as a universal basic income and tax distribution systems will determine whether this optimistic prediction can come true.
(Musk brings up “universal high income” again: when AI takes jobs, government sending money is the solution)
This article The economists highlight job opportunities after an AI layoff wave: the value of scarcity shifts toward “emotional services” first appeared on Chain News ABMedia.
