Artificial Intelligence—Can Japan lead the way?

By Richard Jolley for The Journal (ACCJ)

There are plenty of examples of sci-fi writers imagining products that eventually become commonplace. Think video calling, touch-screen computers, and earbud headphones, the latter imagined by Ray Bradbury in his 1950s classic novel Fahrenheit 451.

On the face of it, the sci-fi genre seems to be a good diviner of things to come, as well as being a popular kind in literature. And we can imagine dedicated research and development departments of companies and governmental divisions reading the latest sci-fi in an effort to peer into the future and outmaneuver rivals.

Given recent concerns about the success, or otherwise, of Abenomics, the name sometimes given to Prime Minister Shinzo Abe’s efforts to revive the economy, it wouldn’t be surprising if the government of Japan had a sci-fi reading department of its own.

It’s clear the economy needs a boost, following decades of stagnation. Data on industrial production at the end of September signaled a second consecutive quarter of negative growth, according to the Financial Times.

Moreover, Abenomics is rapidly running out of “arrows” to aim at the economy, with the third arrow—structural reform—seemingly beginning to drag.

Indeed, a sci-fi ministry isn’t so far fetched when you consider how important the government sees technology as helping solve the country’s woes.

Abenomics puts the Internet of Things (IOT), big data, robotics, and artificial intelligence (AI) at the heart of its revitalization strategy. The goal is to accelerate development across all four fields, and formulate a vision for both the public and private sectors.


Of the four fields, Japan is strongest in robotics. SoftBank’s Pepper — described as the world’s first personal robot — grabbed the headlines when it went on sale here. A lot was made of Pepper’s ability to read a person’s emotions and respond with words and actions that reflected a person’s moods. The product sold out as soon as it became available, with people eager to have a robot of their own.

Yet Pepper still feels a long way from having the kind of intelligence we can expect in the future. For example, it wouldn’t pass the Turing Test, a benchmark in the science of AI for determining whether a machine exhibits behavior indistinguishable from that of a human.

The launch of Pepper raises some important points. On the one hand, it demonstrates Japan’s robotics heritage, accumulated after years of successfully building industrial robots.

On the other, however, it indicates relative weaknesses in areas such as IOT, big data and AI. After all, what have we seen in these areas to compete with the Pepper launch?

Mitsuru Ishizuka is a professor at Waseda University and professor emeritus at the University of Tokyo. He has been working in AI at the latter institution since 1980.

Mitsuru admits that Japan is considerably behind the United States in “deep learning,” a central technology in AI, although the country is working hard to catch up. For instance, the government has established a new AI research center, and is planning another one to promote AI in the country.

As he points out, big data is key for deep learning, and Japan lacks the expertise and financial strength of US multinationals such as Google Inc. or Facebook, Inc. “These companies can invest big money in AI and add the resulting new values to their services. In Japan, there are much smaller companies with specific AI technologies,” Mitsuru explains.

ARTIFiCIAL VERSUS NATURAL The United States has been big in AI for some time now, and the list of companies with IOT, big data, and AI expertise goes way beyond Google and Facebook.

One of the best known is IBM, which developed Watson, the AI computer. Famously, Watson was built to win the U.S. general-knowledge quiz show Jeopardy!, where contestants are presented with answers and must propose the questions.

In 2011, Watson won the show and, since then, has morphed into multiple applications. The computer follows a framework that reflects human decision-making (observe, interpret, evaluate, decide) but can crunch more data than humans ever could.

Japanese entrepreneur and venture capitalist William Saito is a fan. Saito, a professionally trained cook, has used the IBM Chef Watson service to help prepare some unique dishes.

He says: “Designers told Watson to read all the cook books available and then use big data to find patterns of combinations to make something appealing.”

On the whole, Saito’s dishes have been a success, and he can see some exciting applications for Watson’s IOT, big data, and AI mix. “Combine Watson with a refrigerator, for instance,” he says. “You go to your refrigerator and it gives you a recipe based on the food in the fridge prioritized by expiration date.”


Through his work, Saito can distinguish between IOT, big data, robotics and AI in Japan and in the United States. He thinks that AI is seen through a particularly narrow lens in Japan, where too much emphasis is placed on robotic appendages.

The idea of creating cyborgs — humans with mechanical parts — seems to dominate a lot of thinking. This mindset is understandable when you consider Japan has a rapidly ageing population where the demand for assisted living is only going to grow.

But still, it’s a design choice from among many options. Take cars for example. Toyota was in the news recently for financing a $50 million research project on developing an intelligent automobile, as reported in the New York Times.

The project is a collaboration with Stanford University and the Massachusetts Institute of Technology. What makes it interesting is that Toyota is taking a different tack from Google and Tesla Motors, which are developing automated cars. With Toyota, the onus is on the automobile to assist the driver for a safer journey.


For Japan to broaden its horizons around technology, it’s going to require a shift in education, Saito believes. Greater emphasis has to be put on creative thinking in schools, colleges, and universities.

Saito wrote in a recent article for The Journal about Japan’s Galapagos syndrome. He would like to see a better balance between logic and creative thinking, with psychology and physiology sitting alongside traditional computer science and mathematics.

He says this is crucial since IOT, big data, AI, and robotics will increasingly merge, demanding different skill sets. Saito says: “Robotics today in Japan is still very electro-mechanical. It is logically orientated. Robots will need to learn by themselves. Robotics 20 years from now will look nothing like it does today.”


Mazakazu Hirokawa is an assistant professor in the AI laboratory at the University of Tsukuba. He’s seeing the merger of robotics and AI firsthand, and expects the trend to continue as described by Saito.

Hirokawa agrees that the Japanese model focuses more on technology that addresses social issues and is less about creating global solutions. In his work, he focuses on robotics. His ambition is to develop the software to enable a robot to adapt itself to each user and become a “buddy.”

“I’m trying to create algorithms that help robots learn and predictively determine what and how humans want them to act through experience-based inferences,” Hirokawa explains. His aims go way beyond SoftBank’s Pepper, though, to a place where the barriers between humans and machines, not just robots, disappear. It’s a world where the kind of powered full-body suits worn in Iron Man comics aren’t just worn on film sets and at Halloween, but all year round and in everyday life.

It could be some years yet before these kinds of things become everyday, but it’s just a question of time — and software. “We have the hardware to be able to do it, but the important thing is developing the software,” Hirokawa says. “That is the challenge, and I’m working on it.”

Custom Media publishes The Journal for the American Chamber of Commerce in Japan.

© Japan Today

©2022 GPlusMedia Inc.

Login to comment

Artificial Intelligence—Can Japan lead the way?

Lets hope so...there seems to be a lack of natural intelligence in so many areas....

6 ( +7 / -1 )

I didn't even read the article, the headliner is too spot on I'm afraid I'd ruin it for myself. I'm gunna go on with my day with a chuckle.

3 ( +3 / -0 )

For Japan to broaden its horizons around technology, it’s going to require a shift in education, Saito believes. Greater emphasis has to be put on creative thinking in schools, colleges, and universities.

This makes complete sense, but how can it happen if most leaders in academic institutions are so old and stuck in their ways that they are unable or unwilling to understand current technology as well as understand approaches to teaching and learning that differ from what they have always accepted?

0 ( +0 / -0 )

Greater emphasis has to be put on creative thinking in schools, colleges, and universities.

Whoops, let's try another category.

0 ( +0 / -0 )

how a country leads a way if its companies making investments in other countries. yes i am talking to you toyota.

-1 ( +0 / -1 )

Isn't the Japanese government already full of artificial intelligence?

0 ( +1 / -1 )

Japan is building its next artificial intelligence research center in Silicon Valley, California.

1 ( +1 / -0 )

I seriously question the writer's credibility on the given topic.

The launch of Pepper raises some important points. On the one hand, it demonstrates Japan’s robotics heritage, accumulated after years of successfully building industrial robots.

'Pepper' was the brainchild of Aldebaran, A Paris-based company who've been in robotics for over 10 years. Softbank simply acquired them and launched its iteration called Pepper. Its software was not developed in Japan.

the government has established a new AI research center, and is planning another one to promote AI in the country.

"Can Japan lead the way?" - Well, there's your answer. Japan does & will continue to excel in automation & industrial machinery. Without young engineers in a wealthy, progressive company (ie. not Japan Inc.) - never going to happen.

The key factor in developing AI is metadata, and no one can come close to the likes of IBM & Google. They have unlimited budgets & dedicated departments - not to mention scores of acquisitions.

2 ( +2 / -0 )

A quick trip to the ATM will reveal the answer. Poor interfaces, illogical sequences, and user unfriendliness lead one to conclude that the programmers here don't have a clue as to how humans think...

3 ( +4 / -1 )

Not just ATMs. Japanese programmers in general are horrible at logical structures to user interfaces. Rarely do you see an intuitive user interface in Japanese software.

4 ( +4 / -0 )

Japan doesn't even know how to read the map here... None of the big AI players come from Japan, that's sure...

0 ( +0 / -0 )

Dan Lewis, you beat me to it. However, politicians seem to act more out of convenience than anything else.

0 ( +0 / -0 )

Login to leave a comment

Facebook users

Use your Facebook account to login or register with JapanToday. By doing so, you will also receive an email inviting you to receive our news alerts.

Facebook Connect

Login with your JapanToday account

User registration

Articles, Offers & Useful Resources

A mix of what's trending on our other sites