AI Era's Distribution War: Who Wins?

A human and an android went head-to-head in parcel sorting. Figure 03, made by U.S.-based Figure AI, fell behind with clumsy hand movements, dropping items and making mistakes. Humans won.

But after the 10-hour contest ended and the human left complaining of arm pain, the robot kept working. The company plans to keep the robot working until it breaks down as part of the test. Figure 03's labor, which began on the 14th, is being livestreamed on YouTube even at this very moment.

Exactly 10 years ago, AlphaGo defeated Lee Se-dol, marking artificial intelligence's entry into the human intellectual realm. The question of "AI beat a human at Go—so what?" turned into amazement as generative AI models that have since emerged showcased their capabilities. These systems began encroaching on white-collar jobs first. AI's massive data processing and reasoning capabilities, advancing day by day, are overwhelming even high-paid, high-skilled professionals. AlphaFold solves protein structure prediction—work that would take dozens of PhDs a lifetime—in just a few hours. That is why people now say, "The longer you sit in front of a computer, the more likely your job will be replaced."

Now even physical labor, once thought to be a uniquely human domain, is starting to be replaced. Androids like Figure 03 and Boston Dynamics' Atlas are still clumsy. They make mistakes, get lost, wobble and fall. But technological progress will accelerate going forward. Recalling Bill Gates' insight that "people always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten," it is obvious that the inviolable human domain will shrink faster than expected.

Androids are clumsy but unsettling. They don't eat, sleep, or rest. And they're cheap. Atlas is expected to be priced lower than two years' annual salary of a U.S. production worker (about 200 million won). Figure 03 is priced around $20,000 (30 million won). Elon Musk has also pledged to lower the price of his android "Optimus" to $20,000.

The dark factory landscape once seen only in movies is becoming reality. Hyundai Motor announced plans to deploy 25,000 units on production lines at its U.S. auto plants by the year after next. Figure 02 has already gained work experience at BMW's U.S. plant, and Optimus has been deployed as an "apprentice" at Tesla's Gigafactory.

The notion that machines will take on menial work while humans handle high-value tasks is naive. Corporate investment in automation ultimately aims to cut costs. The more expensive the labor, the greater the pressure for AI replacement. Nobel Prize-winning economist Daron Acemoglu has pointed out that "the higher-paid a worker's tasks are, the greater the incentive for companies to automate them." The value of high-paid but "ambiguous" high-skilled workers is likely to fall, while only the value of a tiny core workforce that wields AI freely and maximizes the company's overall productivity will rise further.

In an era when a massive shift in the value of human labor has begun, another unfamiliar scene unfolded on the other side. Samsung Electronics workers, using strikes as a weapon, secured performance bonuses of up to 600 million won per person (based on a 100 million won annual salary). Despite being called "performance" bonuses, even employees in loss-making divisions are expected to take home hundreds of millions of won. In the face of demands amounting to "the company made money, so hand over a lump sum now," both the company and the government—gripped by concerns over disruptions in semiconductor production, the backbone of the national economy—have effectively backed down. If a compensation system in which employees split N% of company profits takes root at Korea's flagship corporation, the repercussions will be far from small. It is likely to set off a chain reaction stimulating competitors, suppliers, and labor unions at other large corporations. There are even concerns that this may signal the dawn of an era of universal struggle.

How long can companies bound by rigid labor regulations keep pouring huge performance bonuses on "ambiguous" high-skilled employees? Or rather, can they survive while maintaining global competitiveness under such a structure? Companies will increasingly want to retain core personnel who maximize productivity through compensation tied to genuine performance, while replacing other personnel with AI.

What is more concerning is what comes next. An era may come in which jobs move not to Korea—where AI adoption is slow due to companies watching out for unions—but to countries that aggressively push AI and automation, like the United States and China. In the past, companies relocated production bases to Southeast Asia in search of low-skilled, low-wage labor. Aren't we underestimating the possibility of an era in which capital moves toward "high-skilled, low-cost, dispute-free" AI labor?