AI (artificial intelligence) is becoming familiar to us, from media to e-commerce to products such as household appliances. On the other hand, there seems to be a difference in the use of AI in business according to the companies. Where is the postman? Providing “Prediction One” AI predictive analytics tool for business sites to more than 20,000 companies in total with Mr. Ryuji Noguchi, who is the author of “Becoming an AI Human Resource in Science – no need for statistics and program knowledge” and belongs to the Japan Deep Learning Association We spoke to Shingo Takamatsu of Sony Network Communications.
As the use of AI grows, few companies can use it for business
――It has been about two years since Mr. Noguchi published “Becoming an AI Human Resources in the Literature-No Statistics / Program Knowledge Required”. Please let us know if you think the recent spread of AI meets or exceeds expectations.
Mr. NoguchiIt is progressing faster than expected. Major platforms such as GAFA (Google, Apple, Meta Platforms, Amazon.com) now offer a number of AI-powered services. Many companies now have AI at the heart of their business.
Even in Japan, AI is used by various companies. At the manufacturing site, we introduced image recognition AI for quality control. Even in autonomous driving, recognition of pedestrians, traffic lights, lanes, etc. is essential. On the other hand, as the number of consumers and users who benefit from AI increases, the gap between the companies that provide and use it and those that do not is widening rapidly.
――While AI is spreading more and more in society, some companies have yet to fully utilize AI. Where is the postman?
Mr. NoguchiThere are several reasons. The first is the question of “people”. Not to mention that there are no human resources who have the knowledge to introduce AI, and there are no human resources who can think about where to use AI in the first place. Moreover, even if AI is introduced, some businesses will fail due to “organizational” issues. I often hear that young employees were rounded up and tried under the order of management, “Do something with AI”, but it ended in a PoC (proof of concept to confirm feasibility) , and the results are proved. does not reach.
–Introduction to AI
Mr. NoguchiEven companies in traditional industries see opportunities to introduce AI. Although it is difficult to create new services using AI, it is possible to replace the work that people were doing with AI.
For example, by streamlining the inefficient work done by human wave tactics with AI, humans will be able to transition to high-value, grassroots work. Moreover, new areas such as natural language AI are emerging, so some companies should seize new opportunities by promoting the introduction of AI.
More important than “understanding AI programs”
――Some large companies hire human resources or create new organizations for the introduction of AI. What are the key points to connect AI to results?
Mr. NoguchiWhen it comes to introducing AI, many people think they need to hire “AI science human resources” such as data scientists and AI engineers. There are also places where we actually hire these human resources or create specialized departments, but it is our role to deal with internal requests such as “I want you to do a little research” and “Please plan this”. There are also places. Even so, there are phenomena that the commercial department queues and does not easily, and the results are slow to come out.
As I wrote in my book, the important thing is to “determine what value will be created using AI and move quickly towards it”. It’s best to make it small, try right away, and if it’s wrong, fix it and try again. In this regard, it is more important for AI HR to understand their own business than to understand AI programs.
–I would like to ask Mr. Takamatsu. Sony Network Communications has provided “Prediction One” AI prediction analysis tool to more than 20,000 companies in total. What do you think of the spread of AI in business?
Mr. TakamatsuI have the impression that this is spreading from large companies to small and medium-sized companies. A growing number of companies are adopting “Prediction One”. I think the key to more popularization is knowing how to use it in the field without being technology savvy.
As more businesses embrace AI, some face challenges. For example, when a system is built from scratch (a method of developing a system from scratch), there are cases where it is difficult to ensure quality. In recent years, AI automation tools have also emerged, and people with AI expertise are also using commercialized tools. The reason is that compared with scratch development, the quality of the released tools is stable, and it is easy to improve the work efficiency of the user service.
In this regard, “Prediction One” has a major feature of a user interface that can be used without specialized knowledge such as machine learning and programming, and can perform highly accurate predictive analysis with just a few clicks. You can easily predict and analyze what the employees on site thought: “Isn’t it possible to use it for this?”
— According to a survey by the Ministry of Economy, Trade and Industry, it is said that there will be a shortage of 124,000 AI human resources in 2030, and the development of human resources is an urgent task . Please tell us about the efforts Mr. Noguchi is involved in with regards to AI human resource development.
Mr. NoguchiIn July 2021, Z Holdings (HD), which owns Yahoo Japan, LINE, ZOZO, etc., launched the “Z AI Academia” program to develop AI human resources across the group. ZHD plans to increase the number of AI employees in the group by 5,000 by 2025. I am also a board member of “Z AI Academia”.
Not limited to AI HR, what is important in HR development is creating a place to gather and learn together. First, when you take a step, the fog seems to dissipate. In addition, by acquiring successful experiences, you will be motivated to learn. I hope it will lead to a successful small experiment at first, and then to a successful experiment that will change society.
Create an environment where on-site employees can use AI on a daily basis
――Sony Network Communications provides “Prediction One” not only to commercial sites, but also to educational institutions and corporate human resource development sites. How should companies go about developing human resources in AI?
Mr. TakamatsuIt’s hard to get established just by going through classroom training. Therefore, by having them use AI tools in training, they can get a concrete picture of AI. Also, if similar AI tools can be used in the field, we should be able to concretely consider the use of AI while using data in the field.
Mr. Noguchi“Prediction One” is also used as teaching material for “Z AI Academia”. I think it’s a great feature that even beginners who have never touched AI can easily handle it.
–M. Noguchi is also a member of the Human Resources Development Committee of the Japan Deep Learning Association. Finally, what are your expectations regarding the development of human resources in AI in Japanese companies?
Mr. NoguchiThe Deep Learning Association conducts an AI certification test called G-test, and this G is a “generalist” G. I feel like we’re in a full-fledged era where generalists are using AI.
I think whether deep learning such as AI can be used in business or not will greatly affect business growth, and even if the labor force shrinks due to declining birth rate and aging of the population, it will create an advantage for the enterprises. you can do it.
――For this reason, it seems that the use of tools will be important.
Mr. TakamatsuIdeally, we should create an environment where everyone can get in touch with AI tools and use them in a practical way, like Excel and PowerPoint.
For this reason, not for AI professionals but for general usersMadeCommercialized AI tools are expected to play an important role. In addition to the AI predictive analysis tool “Prediction One”, we offer services that help solve problems when introducing AI, such as learning how to use it at the stage of introduction, problem definition and data preparation. We are confident that it will be useful not only for companies that are planning to use AI recently, but also for companies that want to use AI more than ever.