The growth of AI in Japan by Rajeev Dutt

Japan is the third-largest economy with the third largest manufacturing output totaling about $1T in exports. Through the 1990s and 2000s, Japan experienced a period of economic slowdown and deflation. In the last ten years, Japan has seen its growth in labor productivity remain stagnant with declining productivity growth in the last three years. Moreover, the working population, aged 15 – 64 has dropped to below 60%. Low productivity growth and a decline in the working population does not bode well for the future economic outlook of Japan.

Against this backdrop, Japanese businesses are looking for ways to compete against the US and China and business leaders are looking at AI. Preferred Networks a Japanese AI solutions company, founded in 2014 and now worth $2B, is a symbol of Japan’s push toward AI.  Preferred Networks is one of the world’s most valuable AI companies and their focus has been on AI solutions. Indeed, Japanese firms are less interested in looking for clever technologies but in solutions that can solve real-world challenges.

Although the need is recognized, and there are some success stories like Preferred Networks, Japan is trailing the US and China in the adoption of AI. There are over 300 AI-related companies in Japan today, but Japan is still lagging. This is a fact that is recognized by business leaders and the government. Masayoshi Son, the founder of Softbank and one of the world’s largest innovation investors, said “Until just recently, Japan was at the cutting edge of global technology. But within the space of a few years, Japan has now completely become a developing country in the most important area of AI where technology revolution is happening.”[1]

Japan has traditionally taken the long view of technology, anticipating significant shifts in technology ahead of the rest of the world. For example, Japan has led the world in the use of statistical process control (ex. Kaizen, a method of continuous improvement created and used by Toyota), innovations in consumer electronics, adoption of robotics, connectivity (mobile internet was introduced to Japan as early as 1999), and contactless payments, QR codes & NFC. Now AI is taking hold in Japan. The concept of “Society 5.0” is being pushed from the highest levels of government to the boardrooms of Japan.

What is “Society 5.0”? If we look at our history, we see an evolution from hunter/gatherer, agrarian, industrial, information, and now AI. Society 5.0 is about making society smart.  This means better governance, healthcare, safety, and security. As Japanese society ages, the need for AI is growing. Already some Japanese retirement homes feature robots and AI monitors that can detect when a resident falls. A retiring workforce means losing highly experienced personnel that cannot easily be replaced. AI can be used in supervisory tasks. A retiring workforce also means fewer people to work in factories or perform services – automation is the only way to ensure that productivity levels are maintained.

That Japan is looking at a wholesale transformation of their society is hardly surprising. Japanese society has long been comfortable with the idea of AI.  Unlike Western societies, where the prevailing archetype of intelligent machines is that of a malevolent force (Skynet and HAL 9000), or where AI is the subject of extensive handwringing by prominent scientists and innovators like Stephen Hawking and Elon Musk, Japanese thinking is driven by the narrative that intelligent machines can be helpful and useful – such as the bioroids in the anime movie Appleseed, or the Tachikoma in the Ghost in the Shell franchise.

Rajeev Dutt walking in Tokyo office at Shono iPark

When Japan faced challenges, it was able to change direction with startling speed. The transformation from a feudal society to an industrial society happened within a generation during the Meiji Restoration. During the late Edo period, Japan had fallen behind the Western world. This became painfully obvious when US Commodore Matthew Parry came to Japan in advanced, heavily armed warships in 1853 to coerce the Japanese government to open up their ports to trade.  In 1867, Emperor Meiji ascended to the throne and ushered in a market-based industrial society to catch up to the West.

After the second world war, Japanese industry was in tatters. As the country started to recover, there were severe quality issues with Japanese products. “Made in Japan” was a pejorative, indicating shoddy products and poor quality. Once again, Japan rose to the challenge. In 1950, W. Edwards Deming, an American engineer, and statistician introduced the “Statistical Product Quality Administration”, which led to the Japanese post-war economic miracle in the 1950s and 1960s.

Today, we can see the same patterns repeat. Unlike the US and China, Japan simply does not have the deep talent pool of machine learning engineers and scientists; nevertheless, Japanese businesses are beginning to see that their dominance is eroding. To short-circuit the path to adoption, Japanese businesses are increasingly looking at platform-based solutions that offer a single, cohesive approach to the creation, deployment, and management of AI. They are looking at ways to transform their troves of data into real intelligence whether in the back office or on the factory floor.

Moreover, Japanese businesses want to have these AI platforms on-premise. Japanese businesses are wary of pushing their data into the cloud for several reasons. The first is the question of confidentiality. Asian companies, in general, prefer that their data remain on-premise or within the company’s direct control. The idea of pushing their data into the cloud is met with skepticism. The second reason is that data resides on-premise and it is difficult, expensive, or even, in some cases, impossible to move into the cloud. To understand why one must look at the composition of the Japanese economy. Manufacturing consists of 19% of the Japanese economy, compared to just 10% of the US economy. For business-critical operations such as manufacturing, a dependency on the cloud has two handicaps – the first is poor internet connectivity and the second is the fact that latency often exceeds the threshold of tolerance for many processes. Decisions must be made in the order of milliseconds and not hundreds or thousands of milliseconds it would take to move the data into the cloud and receive an answer from a cloud-hosted AI.

The move to on-premise platforms or solutions means that the AI market in Japan is dominated by either professional services offered by companies like Preferred Networks or Accenture or on-premise platform companies such as DataRobot or c3.ai. There is also a growing acceptance of startups in this domain simply because of the lack of options on the AI platform side.

Rajeev Dutt walking at the office at Shonan iPark, Japan

It is in this environment that AI Dynamics is finding increased traction in Japan. Our two focus markets are Manufacturing and Biotechnology. Manufacturing is undergoing a significant shift in thinking as the next generation of technology is transforming factories into smart factories that are capable of self-optimization, managing supply chains on their own, or even supervising employees – all with the use of artificial intelligence. This shift broadly comes under the moniker of “Industry 4.0”. Japanese industries are keen to start adopting these technologies, not because they will yield benefits in the next year or so, but because in five to ten years, they anticipate that productivity challenges will necessitate these changes.

Japanese pharmaceutical companies are also looking at driving innovation in drug development and the creation of new therapeutics.  Takeda, Asia’s largest pharmaceutical company, and one of its oldest (it was founded about 239 years ago) was once in the top ten pharmaceutical companies globally. Today it is in 19th place. Takeda Pharmaceuticals is looking at AI to help design new therapeutics, optimize drug development, identify biomarkers, improve drug safety, and aid in manufacturing.

Our first subsidiary outside of the US is in Japan. As a platform company, we are seeing the greatest interest in Japan and Asia more broadly. If history is any guide, Japan is in the midst of another major transformation – to Society 5.0 and we believe that we are in the right place at the right time to participate in this incredible opportunity.


[1] https://www.analyticsinsight.net/japan-losing-tech-position-lags-adopting-artificial-intelligence/

Leave a Reply

Privacy Preferences
When you visit our website, it may store information through your browser from specific services, usually in form of cookies. Here you can change your privacy preferences. Please note that blocking some types of cookies may impact your experience on our website and the services we offer.
%d bloggers like this: