Algorithm discount fuels human premium

In June 2025, I was traveling through Georgia, a small country in the South Caucasus. At a market in Tbilisi, I stopped at a stall and noticed a painting no larger than my palm. The price was 300 lari ($115). The seller was a woman in her 50s. I could not tell whether she had painted it herself. What made the question harder was the company the painting kept: more than a dozen obvious reproductions, including Mona Lisa, priced at 30 to 80 lari.

Once art can be mass-produced cheaply, the work that carries the trace of a human hand acquires something that machines cannot manufacture on demand: a premium.

Call it the human premium.

When consumers know that a work, product or service was genuinely created by a person, they may be willing to pay more for it — not necessarily because it is objectively better, but because "being human" has itself become a scarce attribute.

A 2025 study in the Journal of Management Information Systems found what its authors call the "algorithm discount": consumers value digital products less when they learn that algorithms, rather than humans, created them. The effect is driven in part by the effort and emotional investment consumers imagine being embedded in human-made work.

This is not entirely new. The Industrial Revolution made machine-produced textiles, furniture and household goods dramatically cheaper.

The Arts and Crafts movement that followed, associated with figures such as William Morris, helped turn workmanship itself into a form of distinction. Luxury goods still depend on much the same logic: scarcity, craftsmanship and the visible involvement of human hands.

Walter Benjamin described a related phenomenon in his 1935 essay The Work of Art in the Age of Mechanical Reproduction. He called it aura — the sense that an original possesses a unique presence tied to a particular time and place. Photography and film, Benjamin argued, weakened that aura by making reproduction effortless.

Generative AI has pushed the argument somewhere Benjamin could not have imagined. AI does not merely reproduce existing works. It can generate new text, images, music and software at almost negligible marginal cost. The result is neither quite an original nor merely a copy. It is something closer to synthesis.

And when synthesis becomes abundant, the human original becomes rare.

That is where aura acquires an economic meaning. If technology drives the cost of what a capability produces toward zero, the capability itself tends to lose economic value.

What technology cannot easily reproduce may move in the opposite direction.

The machine revolution once revived the value of the handmade. The AI revolution may be doing the same for human judgment, taste and authorship.

Publishing offers an early glimpse of what this market could look like. In 2025, the Authors Guild, the oldest and largest professional organization for published writers in the United States, introduced its "Human Authored" certification, allowing eligible books to carry a mark indicating that their text was written by humans, while permitting minimal AI assistance such as spelling or grammar checks.

The certification can be verified through a public database; non-members currently pay $10 per title.

The significance is larger than the badge itself. A market is beginning to build infrastructure around a simple proposition: provenance matters.

And readers appear willing to pay for provenance. Substack reached five million paid subscriptions in March 2025, turning individual writers into a substantial subscription economy.

Its appeal lies in precisely what mass-generated content cannot easily supply: a recognizable person with a distinctive voice, history and point of view.

Yet the human premium is not necessarily good news for everyone.

The more routine creative work AI can perform, the more pressure falls on the middle of the labor market.

Research from Ramp Economics Lab found that companies' spending on freelance marketplaces fell from 0.66 percent of spending in the fourth quarter of 2021 to 0.14 percent in the third quarter of 2025, while spending on AI model providers rose from virtually nothing to nearly 3 percent.

That is the uncomfortable arithmetic of the human premium. AI may not eliminate the value of human work. It may concentrate that value.

A novelist with a recognizable name, a designer with a distinctive aesthetic or a journalist with a trusted reputation may become more valuable precisely because audiences can identify the person behind the work.

But an anonymous writer producing competent, interchangeable copy may discover that competence is no longer enough.

The result could be a sharply divided, K-shaped creative economy: extraordinary premiums at the top, collapsing prices in the middle and an ocean of synthetic content at the bottom.

There is another problem. The moment "human-made" becomes valuable, it becomes something worth faking. The Authors Guild's system relies on identity verification, registration and contractual enforcement rather than a foolproof technological test of authorship.

That is not necessarily a flaw; it is a reminder that the real scarce asset may not be human labor at all, but trust.

The deepest irony is that the human premium depends on AI. Without a flood of machine-generated words, images and music, human creation would not suddenly look so scarce. The premium is, in that sense, a child of abundance.

That may be the defining economic pattern of the AI age. Whenever technology makes something virtually unlimited, the market begins searching for the version that cannot be mass-produced. Scarcity does not disappear; it moves.

For centuries, we paid more for things because they took longer to make. In the coming decades, we may pay more because someone, somewhere, actually cared enough to make them.

The question is no longer whether machines can create. They plainly can.

The more difficult question is what we will continue to value once they can create almost everything.

Perhaps the answer will not be that humans can always do it better.

Perhaps it will be that a human did it at all.

The author is a senior research fellow at the Tencent Research Institute.

The views don't necessarily reflect those of China Daily.

If you have a specific expertise, or would like to share your thought about our stories, then send us your writings at opinion@chinadaily.com.cn, and comment@chinadaily.com.cn.