As the "Fourth Industrial Revolution" boom came amid advances in artificial intelligence technology in the latter half of the 2010s, many countries including Japan formulated AI promotion strategies. Recently, the use of generative AI such as Chat GPT has spread rapidly. An increasing number of companies are using generative AI for documentation, sales activities, customer support, etc.
Since AI is a general-purpose technology with a wide range of applications, its development and spread has the potential to significantly increase the productivity of the economy as a whole. It has been pointed out that if AI reaches a technological singularity where AI exceeds human capabilities, economic growth rate could explode. At the same time, it has been estimated that AI would replace human labor, leading to high unemployment. At present, the technical singularity is seen as a matter of distant future. How large of an effect will AI have on productivity and economic growth in the near future?
From the mid-1990s to the early 2000s, the information technology revolution accelerated productivity growth in the United States and in several other countries. The contribution of IT-using industries such as finance, transportation, and retailing to macroeconomic productivity growth was greater than that of IT producing industries such as computer manufacturing and software, because IT-using industries are larger than the IT producing industry. While research on the effects of automation technologies such as industrial robots on productivity and employment has progressed rapidly, empirical analysis of AI’s economic effects has been lagging. The main reason for this is the lack of comprehensive data on the use of AI by industries and companies.
According to a survey of medium-sized and larger Japanese companies I conducted at the Research Institute of Economy, Trade and Industry, only 3% of companies were using AI five years ago. A recent follow-up survey found that 10% are now using AI. In particular, companies with a higher proportion of highly educated employees with university or graduate school degrees are more likely to use AI. Of companies that use AI, 80% expect positive effects of AI use on their long-term productivity.
A recent survey of Japanese workers shows that about 6% use AI for work. Highly educated people are more likely to use AI, suggesting that cognitive skills and AI are complementary at present. About two-thirds of AI users reported an increase in work efficiency. The average productivity growth surpassed 20%, although their growth estimates are highly dispersed.
Some studies have been conducted to rigorously measure the effects of AI through randomized experiments targeting specific tasks such as writing, programming, and customer support. Although the sizes of the effects differed widely from task to task, the effect of increasing productivity was confirmed for the majority of them.
Since it is certain that more people will use AI for work in the future, AI is expected to have a significant positive effect on the productivity of the economy as a whole. However, it should be noted that there is a selection effect whereby AI is used more by people who are engaged in tasks for which AI has greater positive effects on productivity. This means that the existence of industries and occupations to which it is difficult to apply AI may constrain macroeconomic effects. In order to increase AI’s economic effects, AI application may have to be expanded to a wide range of tasks beyond the boundary of white-collar desk work.
>> Original text in Japanese
* Translated by RIETI.
May 9, 2024 - Published in Nihon Keizai Shimbun's "Economist 360° Perspective"
