The "2022 AI Predictions Survey Japan Edition" has been released by PwC. The survey reveals the current state of AI utilization by Japanese and U.S. companies, as well as where both companies are making progress and where they are lagging behind. What should Japanese companies do to start using AI?
June 01, 2022 07:00 Published.
[Ikuyo Yoshida, ITmedia]
This article is for members only. Please register to view the entire article.
PwC Consulting (PwC) released on May 19, 2022 its "2022 AI Prediction Survey: Japan Version," comparing the use of AI by Japanese and U.S. companies.
The survey compiles the results of a web-based survey conducted in Japan and the U.S. in January 2022. In Japan, the survey covered 300 respondents at the managerial level or above at companies with sales of 50 billion yen or more that have already introduced AI or are in the process of considering its introduction. In the U.S., the survey covered 1,000 executives at companies with sales of $500 million or more that had implemented AI or were in the process of considering its implementation.
Japanese companies (53%) vs. U.S. companies (55%)
First, the survey asked about the current status of AI adoption by companies. 53% of Japanese companies and 55% of U.S. companies responded that they are "AI adopters," with a close gap of 53% in Japan and 55% in the U.S. In addition, 11% of respondents in Japan and 18% in the U.S. said they were "preparing to introduce AI," while 36% in Japan and 27% in the U.S. said they had not yet introduced AI.
Comparison of the degree of AI utilization by Japanese and U.S. companies (Source: PwC presentation)
Hiroyuki Nakayama of PwC (AI Lab Leader, Data & Analytics Leader, PwC Japan Group) said, "There are two hypotheses as to why there has been no significant progress in AI utilization by US companies: First, not all companies, even in the US, are AI advanced companies, and about 60% of them are not AI advanced companies. The first is that not all U.S. companies, even U.S. companies, are AI advanced companies, and that about 60% of them, including their followers, may have reached a certain "ceiling". The other hypothesis is that U.S. companies tend to seek short-term results and that some projects have been halted due to the economic stagnation caused by the Corona disaster. We plan to keep a close eye on 2023 and beyond to see which is closer to the truth," he says.
What is lagging behind the U.S. is their ability to measure ROI and the degree to which they have implemented AI governance.
Although it is largely due to their footsteps that they have caught up with the U.S., it is also true that the number of companies using AI in Japan is increasing. This is due in part to the "DX Report" published by the Ministry of Economy, Trade and Industry (METI), which has increased the momentum of digital transformation (DX) among Japanese companies, as well as to the unique Japanese characteristic of "many followers following a few innovators. Mr. Nakayama believes that these factors are also related to the unique characteristics of Japan, where "many followers move in pursuit of a few innovators.
However, it is not all smooth sailing in Japan either. One of the areas in which Japan lags behind the U.S. is the ability to measure ROI (return on investment). According to the survey, 64% of companies in the U.S. are able to accurately measure the ROI of their AI investments, compared to only 21% in Japan. U.S. firms also outperformed U.S. firms in all categories in the ratio of implementation of various AI governance measures.
Comparison of the ratios of various AI governance measures implemented by U.S. and Japanese companies (Source: PwC presentation)
Takuya Fujikawa, Data & Analytics Leader at PwC, commented, "In AI applications, it is important to start with a proof-of-concept (PoC) and, once results are achieved, to deploy it across the company and reap the benefits. In terms of AI governance, the graph shows that U.S. companies are not only establishing rules, but also taking action on them. Japanese companies have been studying the issue, but have yet to take action," says Fujikawa.
According to Mr. Fujikawa, it is essential for companies to take the initiative in AI utilization from the perspectives of "in-house production" and "self-driven. While companies that are making progress in AI utilization are moving toward in-house production in areas such as "theme creation and planning," there are some phases that are not yet in the spotlight. That is "operational improvement (MLOps). While 47% of the companies cited "theme creation/planning" as a phase they would like to work on in 2022, only 23% cited "operation improvement" as a phase they would like to work on in 2022.
Low attention to in-house production in the operation and maintenance phase (Source: PwC's presentation material)
The phase that maximizes the effectiveness of AI utilization is the operational improvement phase, since AI utilization is a process of finding best practices through repeated trial and error. The most important thing is to bring AI in-house, and I believe that the current lack of attention is a major problem for AI utilization in Japanese companies," said Fujikawa.
New Trends of Fusion with Other Technologies and Data Distribution
The survey also revealed new trends in the use of AI by Japanese companies. Among them, Fujikawa and his team are focusing on the "fusion of AI and other technologies." Now that the use of AI on its own has come to a halt, there is a trend to combine AI with other technologies to solve new problems.
The technologies to be combined include IoT, robotics (excluding RPA), augmented reality (AR), virtual reality (VR)/metaverse, 3D printing, blockchain, drones, quantum computers, and many others. More than 50% of companies in all categories were found to be in "currently considering" or higher status.
Companies considering the use of integration with diverse technologies (Source: PwC's presentation).
In addition to this, "data distribution" is another trend. Data is being distributed not only internally but also externally, and conversely, external data is being used internally. From this survey, when the status of "currently considering" is included, as many as 70% of Japanese companies were already working on this theme.
Maximizing Effectiveness by Measuring ROI and Aiming for Socially Acceptable AI Application
Based on the results of this survey, what advice would you give to Japanese companies that are planning to use AI in the future?
We would like them to make sure to measure the ROI of their PoC projects. Japanese companies will be required to maximize the effect of their efforts based on objective evaluations. In addition, AI will further expand in the future and become a social infrastructure. In order to use AI in such a way that it will be accepted by society, it will be important for companies to take a proper approach to AI governance and to announce their efforts to the public in order to provide customers with peace of mind and safety.
AI investment is still in its early stages, and there is a risk that if ROI is strictly measured, it will not be possible to move forward, but Mr. Nakayama says that it is still necessary to look back.
However, Mr. Nakayama said that it is still necessary to look back. "Some U.S. companies have hypothesized that measuring ROI has prevented them from moving forward with AI investment. However, I think it is important to look back on the implementation of AI because it is not going as smoothly as we would like. It is important to think about whether or not the implementation was meaningful, and since it is impossible to get a score of 100 from the beginning, it is important to go through the process thoroughly while learning from it. For this reason, I think it is important to measure ROI as a way of looking back," said Nakayama.
Copyright © ITmedia, Inc.
0 コメント:
コメントを投稿