An ice cream developed by a Japanese dessert The reason why Japan is slow in developing generative AI is…Lack of engineers and supercomputers”

(Source from Reuters/Alamy)

Since the advent of the generative artificial intelligence (AI) service “Chat GPT,” AI craze has been blowing around the world. Governments and companies around the world, including the U.S. and China, are scrambling for generative AI. However, Japan, which has led the high-tech market in the past, is relatively quiet.CNBC, a U.S. economic media outlet, said, “Japan is lagging behind in the generative AI market,” citing a lack of software engineers to take charge of the LLM (Large Language Model), the core of AI development, and supercomputers to perform it.AI-related signs installed at the trade fair ‘Next Tech Week Tokyo 2023’ held in Tokyo, Japan on May 10. LLM development is actively taking place in the U.S. and China. China’s Alibaba and Tencent have introduced at least 79 LLMs in China over the past three years. In the U.S., Google and Microsoft (MS), as well as open AI that released Chat GPT, are leading LLM development.On the other hand, Japan lags behind the United States, China, and Europe in terms of the size and speed of LLM development. First of all, there is a lack of software developers to develop LLM. According to Japan’s Ministry of Economy, Trade and Industry, 789,000 software developers are expected to be short in Japan by 2030.

(Source from Reuters/Alamy)

Noriyuki Kojima, co-founder of Japanese AI startup “Kotoba Technology,” told CNBC, “Japan’s sluggishness in the generative AI field is mainly due to a relative lack of deep learning and software development.”Supercomputers such as IBM’s “Bella” are needed to develop LLM, but Japan is also lacking. Nikkei (Nihon Keizai Shimbun) pointed out in May that “high-performance supercomputers are needed to build LLM, but none of the Japanese companies have world-class supercomputers.” The Japanese government plans to invest 6.8 billion yen to install supercomputers in Hokkaido as early as next year.Of course, there are companies in Japan that are interested in generative AI. ARM, a British semiconductor designer owned by Japan’s Softbank, announced in June that it plans to develop its own generative AI platform. NTT, a Japanese telecommunications company, also announced that it will develop its own LLM to create light and efficient services. NTT plans to invest 8 trillion yen in the data center and AI sectors over the next five years.

KS CHOI

US ASIA JOURNAL