The debate that AI will take away jobs has really been everywhere lately, but actually, from an economic perspective, it's completely wrong.

There's a common claim that "the total amount of jobs is fixed," but this completely ignores 200 years of economic history. Let's look at the era when agriculture was mechanized. In the early 20th century, one-third of the U.S. workforce was engaged in farming, but by 2017, that had dropped to 2%. When tractors appeared, some probably thought, "This will cause all agricultural workers to become unemployed." But in reality, agricultural output tripled, and those workers flowed into factories, offices, hospitals, research labs, and the software industry.

Electricity is the same. It wasn't just a switch in energy sources; it transformed entire factory structures and created entirely new consumer products. In the early 20th century, only about 5% of factories used electricity, but by 1930, 80% of manufacturing was powered by electricity. Over the following decades, labor productivity doubled, and employment in manufacturing, sales, and credit provisioning exploded.

When Excel and VisiCalc appeared, it was thought that bookkeeping jobs would disappear. But in reality, while 1 million bookkeepers were laid off, 1.5 million financial analysts were hired. The nature of the work changed, but overall employment expanded.

Here, a crucial rephrasing is needed. It should be understood as "job reorganization," not "job loss." As productivity improves, new needs emerge. When food becomes cheaper, spending increases in other areas like housing, healthcare, education, travel, entertainment, and convenience. This leads to the birth of new industries and new job types.

So, what about AI's current position? Looking at academic research, the results are clear. A paper from the NBER states that "the adoption of AI has not significantly changed overall employment levels." A survey by the Federal Reserve Bank of Atlanta estimates that over 90% of companies saw no impact from AI in the past three years. Data from the Census Bureau shows that only about 5% of companies using AI reported effects on their employment numbers, with increases and decreases roughly equal.

More notably, there’s a rapid increase in hiring software engineers and product managers. From early 2025, jobs in software development have been steadily growing. This indicates that AI is functioning more as an "augmentation" rather than a "replacement." Routine administrative tasks are decreasing, but analytical, technical, and managerial roles are increasing.

According to Goldman Sachs estimates, the "substitution" effect of AI is much smaller than the "augmentation" effect. During earnings calls, the phrase "leveraging AI as an enhancement" is used about eight times more often than "using AI as a substitute."

The emergence of new businesses is also exploding. The proportion of new apps appearing in app stores increased by 60% year over year. Robotics is gaining attention as a completely new industry, with robot-related datasets jumping from 10th to 1st place in two years.

Certainly, some jobs will decline. The U.S. Bureau of Labor Statistics predicts decreases in roles like customer service representatives and medical administrative staff. But that’s not the end. Overall, in the economy, there’s no statistically significant correlation between AI and unemployment or job growth rates.

Ultimately, human needs are endless. The total amount of work and cognitive capacity are not fixed. AI is not the end of jobs but the dawn of a richer era of intelligence. Just as in the past, new industries, new companies, and new roles will continue to emerge.