Surveys reveal AI-driven efficiency fails to reduce working hours or increase employment

Motoki Tamura of the same institute points out that the cause is the fact that the framework of working hours, such as the 40-hour work week, remains unchanged. He suggests that even if tasks are made more efficient by AI, new duties are added, resulting in a system where the total workload remains the same or any excess becomes overtime.

Furthermore, there is a challenge where even if AI reduces the amount of labor (the denominator), if the added value (the numerator) does not increase, the productivity of the company does not substantially improve, even though the calculated productivity increases.

Meanwhile, concerns about the impact of AI on employment are spreading globally. According to a survey of 37 countries worldwide conducted by the Pew Research Center (announced on 2026-09-17), in most countries, including Japan, the proportion of people who responded that "AI will cause job losses" exceeded those who responded that it would "cause job growth." In Australia, South Korea, and the United States in particular, responses predicting job losses exceeded 70%.

The same survey also showed that in many countries, younger generations aged 18 to 34 tend to be more concerned about job losses than older generations aged 50 and above. Additionally, there was an overall tendency to believe that AI will widen the gap between the rich and the poor, a trend that was found to be particularly strong in high-income countries.

PLUS ULTRAby Amenoyomi

Even if individual tasks are shortened by AI, there is a structural issue where new tasks are packed into the freed-up time as long as the total framework of working hours—such as the 40-hour week—remains unchanged. For example, if a job that previously took 8 hours is reduced to 4 hours through the use of AI, that task tends to be redefined as a "job that finishes in 4 hours," and other tasks are assigned to the remaining 4 hours. The Persol Comprehensive Research Institute survey also indicated that while task times were shortened by AI, only 25.4% of users saw a decrease in overall working hours, with many of those who did reduce their time spending it on other work such as routine operations.

This phenomenon can be organized through the concept of labor productivity. Labor productivity is calculated as "Added Value (Numerator) ÷ Amount of Labor (Denominator)," but AI primarily reduces the denominator, the amount of labor. If the time required to produce the same output is halved, the calculated productivity improves; however, it is analyzed that unless the value provided (the numerator) itself increases, it is difficult to link this to actual improvements in corporate productivity or price pass-throughs.

On the other hand, anxiety regarding AI's impact on employment is spreading worldwide. In a survey of 37 countries conducted by the Pew Research Center, the number of people who predicted that "AI will cause job losses" outnumbered those who believed it would "cause job growth" in most countries, including Japan. In Australia, South Korea, and the United States specifically, over 70% of respondents predicted job losses.

These concerns tend to appear more strongly in countries with higher GDP per capita. Furthermore, by age group, it has been reported that anxiety over AI-induced job losses is more common among the youth aged 18 to 34 than among older adults aged 50 and above.