Business | DAVOS SPECIAL 2017

A future of promise and pitfalls as AI technology advances

by Shusuke Murai

Staff Writer

On a Sunday afternoon in 2035, a man in Tokyo is consulting a robot as to what he should eat for dinner.

“It seems that you have a heavy feeling in your stomach today; I suggest fish and vegetables,” the robot said, based on the man’s health data retrieved automatically from a device he is wearing. “I would suggest ordering a cabbage from this farm, as it matches your preference for pesticide residue limits.”

“Okay,” the man says. Soon after the cabbage was delivered from the farm by drone, the robot installed a recipe data file to a cooking machine to prepare dinner for the man.

This is a scene of daily life that the government envisages as the future of Japan’s “super smart society,” where all individuals can receive the necessary services they need regardless of differences in age, gender, region and language.

Japan has long been gearing up for developing information technology and robotics powered by artificial intelligence (AI).

Last year, the government launched the Fifth Basic Plan for Science and Technology, which, by promoting the development of advanced technologies such as the “Internet of Things (IoT),” big data and AI, aims to set the foundation for a smart society by March 2021.

Tokyo believes technological advancement is a much-needed boost for industries in Japan, where the dwindling workforce caused by the falling birth rate and an aging population has been an imminent threat.

In the Japan Revitalization Strategy 2016, the government said it aims to cover the labor shortage by improving the productivity of the industrial world through the paradigm shift brought by advanced technologies like AI and IoT.

Indeed, the expectation is high on the rise of advanced AI, which has achieved some notable milestones in 2016.

In March, the AI-based computer program AlphaGo, developed by Google subsidiary DeepMind, stunned the world when it defeated South Korean go grandmaster Lee Sedol in a five-game match of the ancient board game that requires strong intuition and creative thinking.

And in August, it was learned that Watson, IBM’s artificial intelligence platform, helped doctors rapidly diagnose a rare type of leukemia that conventional tests had failed to detect — possibly the first case in Japan that an AI-powered program contributed to saving a patient’s life.

The present hype over AI, which began early this decade, can be described as the “the third AI boom,” according to Yutaka Matsuo, project associate professor at the University of Tokyo.

The current innovation surrounding AI is largely driven by the rapid growth in a method of machine learning called deep learning, Matsuo, Japan’s leading AI expert, said during an interview with The Japan Times last year.

Thanks to deep-learning technology, which is inspired by how the human brain learns things, artificial intelligence has become able to automatically extract patterns from vast amounts of image data and classify them without being instructed by human beings.

The changes happening around deep learning are “revolutionary,” Matsuo said.

“Let’s say there was a person with a brain disease. The healthy eyes send visual information to the brain, but it cannot recognize what it’s seeing. … That has been the conventional computer,” Matsuo said.

But deep-learning technology provided computers with image recognition ability that even surpasses that possessed by humans, he said, adding that the technology can, for example, be used to detect possible incidents when implemented in security cameras — like identifying drunken drivers or people inclined to start fights in the street.

Although tech giants in the West have led the competition over information technology, Matsuo believes Japan can catch up in the development of deep learning because the innovation happening now is around image processing technology, an area where Japan leads the world.

“In deep learning, Japan still has a great chance to compete, and an excellent opportunity to gain great benefit out of it,” he said.

Meanwhile, some critics fear that smarter AI and robotics may put people out of work.

According to a report published jointly in December 2015 by private think tank Nomura Research Institute and Oxford University, researchers predicted that advanced robots and AI would replace about 49 percent of Japan’s workforce within 10 to 20 years.

The report further explained that jobs that didn’t involve special knowledge or skill — including receptionists, taxi drivers, delivery workers and security guards — would be taken over by machine, while those that require creativity and the ability to understand others — like economists, video game creators, teachers and writers — would remain in human hands.

Some experts, including Noriko Arai, a mathematician at the National Institute of Informatics, believes the chilling future predicted in the report is not far off the mark.

After five years of developing AI that aims to pass the entrance exam of the prestigious University of Tokyo, or Todai, Arai concluded that Japan needs to improve students’ reading comprehension — an area where AI has yet to catch up — before improving AI.

“The worst-case scenario is, from 2021, when the number of people aged 22 or below will drop off significantly, the world will be short of workers, but the number of unemployed will increase as well,” Arai said during an interview with The Japan Times.

As a result of her much-noted Todai Robot Project, the ultimate aim of which was to gauge AI’s possibilities and limits, while predicting the future of an AI-enhanced society, Arai found that AI is good at answering fact-based questions — better than most test takers — because it can take advantage of the vast information stored in its database to find answers.

However, AI failed to come up with correct answers requiring common knowledge not outlined in the question, because it merely output most likely answers after interpreting a question by identifying keywords and phrases. Therefore, the AI cannot think outside the box or have the kind of deep understanding humans have about the meaning of a question.

Despite the limitations in machine’s capabilities, Arai paints a bleak picture of the future because she found out that many students lack ability to think creatively — the thing that gives them an edge over machines.

That is why she last year decided to pause her AI’s quest to enter the University of Tokyo, and instead focus on improving students’ scores on a reading skill test, a multiple choice test developed by the project team to test their ability to visualize an image from a written sentence — essentially to think for themselves — so that they can survive the AI-saturated society.

“The key thing would be whether people are able to exploit machines to supplement their abilities, rather than follow what machines tell them to do,” Arai said.

Arai’s view on how to survive an AI-powered society is also shared by Matsuo, who believes the essence of the change in the society is not about stealing human jobs, but about redefining tasks to allow people to focus on more creative, communicative things rather than the repetitive, menial labor that computers are better at performing.

“The important thing is to continue updating their skills to keep up with changes in their tasks rather than to be satisfied with skills and knowledge they acquired in the first 20 years of their lives,” he said.


Download the PDF of this Davos Special