Although artificial intelligence (AI) is rapidly spreading to corporate sites, there is no clear cha..

44% of U.S. businesses use AI
Introduction Speed When PCs were distributed in 1980, the productivity growth rate was only 2.2%, and the cost of introducing less than
AI offset the productivity effect

Although artificial intelligence (AI) is rapidly spreading to corporate sites, there is no clear change in productivity statistics. The pace of AI introduction by U.S. companies is about three times faster than when personal computers (PCs) were distributed about 40 years ago, but the productivity growth rate is not up to the level at that time. While the theory of controlling the speed of AI development is being raised, there are also questions about the expectation that AI will increase productivity.

According to a Reuters analysis on the 16th (local time), about 44% of U.S. operations were using AI as of May. The figure is based on tracking data on the introduction of Generative AI supported by Harvard Business School.

However, the spread of AI is not directly linked to an increase in labor productivity. In the second quarter of this year, hourly production in the U.S. increased by an annualized rate of 2.2% from the same period last year. It is higher than the annual average growth rate of 1.5% in the 1980s, but less than the productivity growth rate of 2.7% in the late 1990s when PC investment began in earnest. At that time, there were cases where the quarterly productivity growth rate exceeded 4%.

Contrary to expectations that AI will significantly increase economic productivity than PCs, the effect has not yet been clearly shown in statistics. However, there is a possibility that the current statistics do not sufficiently capture the economic value created by AI.

For example, if products designed in the U.S. are produced and sold abroad, such as Nvidia’s AI semiconductor, the effects on the productivity and GDP of the U.S. economy may not be sufficiently reflected in the statistics. Epoch AI, an AI research institute, estimated that the annual GDP measurement of the United States could be up to 2 percentage points lower than it actually is in 2028, given these problems.

However, Reuters pointed out that it is difficult to explain all current productivity trends just by explaining that statistics do not properly reflect the effects of AI. This is because the inflation rate is not much different from the global trend even in the wholesale and financial industries, where the AI introduction rate exceeds 60%. This means that AI may not work as a technology that boosts productivity in a short period of time like PCs during the Internet revolution.

Costs incurred in applying AI to actual tasks by companies are also cited as a hindrance to productivity improvement. Companies are retraining employees and hiring outside consultants to utilize AI. In the process of replacing human resources using AI, new human resources and costs are being put in.

Reuters analyzed that the cost of AI investment, work redesign, and employee training could offset the effect of improving productivity in the short term. It may take time for AI introduction to show tangible changes in productivity statistics.