TOKYO -- Asia is evolving into a popular home of artificial intelligence experts with China leading the way, but Japan has proved to be an outlier as its talent pool falls behind the world's smaller economies in both size and quality.
Canadian startup Element AI found that there were 22,400 people worldwide it regarded as top talent in its annual report on people working in the AI field. Nearly half, 10,295, were in the U.S. China was second with 2,525, followed by the U.K. with 1,475, Germany with 935 and Canada with 815. Japan, despite being the world's third-largest economy, placed sixth at just 805, or 3.6% of the total. The figures were derived from surveying the number of authors who published papers at international conferences in 2018, analyzing their careers and breaking down the distribution by location.
AI research and development is a cutting-edge field led by experts in universities and corporations, while the business side is supported by administrative talent with specialized knowledge. The growing prevalence of "internet of things" technology -- where devices and systems are linked and able to share information -- and big data increasingly make AI expertise crucial in developing products and finding new business opportunities. A dwindling of talent at the top of this field could result in stagnant adoption of AI by companies, deteriorating corporate and national competitiveness in turn.
Japan's Ministry of Economy, Trade and Industry estimates that the nation will face a shortage of 120,000 AI business experts by 2030 as demand for the technology in cars and finance grows.
The quality of each country's talent pool goes beyond sheer size, however. Diversity is also important. Element AI found that 40% of Singapore's experts returned to domestic companies after studying overseas. A high percentage of talent in China and South Korea did the same at 30%, while India rose to 25%.
Women also comprised 20% or more of Taiwan, Singapore and China's top talent, beating the global average.
Japan's talent pool, meanwhile, lacked such diversity. Only 17% studied overseas before returning home to work at a Japanese company, the second fewest of 17 major places surveyed. The country also had the fewest percentage of women at 9%, half the global average. South Korea also filled out the bottom at 12% in terms of female AI talent.
Diversity among experts is crucial to the development of AI. A mixture of people from different backgrounds helps in obtaining cutting-edge technology from around the world and conducting joint research with other organizations. It is also thought that teams composed of similar people tend to make biased decisions since they do not consider a variety of perspectives.
Japan's lagging education system is partially responsible for the gap. China, the U.S. and European countries have made fostering talent in science and technology a matter of national policy.
The U.S. released a plan to revitalize science, technology, engineering and math education over a decade ago, greatly increasing the number of staff in colleges and high schools devoted to those fields and widening their scope. China also launched a plan to develop AI in 2017 and has set up AI departments in schools.
Japan, meanwhile, has maintained the traditional composition of subjects, which has hurt the development of AI talent. Learning the technology requires research across several fields, including math and computers. Japan only has three universities where students can obtain a degree in data science. In the U.S., many schools offer such programs while top schools like Stanford University and the Massachusetts Institute of Technology attract prominent researchers and students from around the world.
"AI development requires mathematical knowledge and ability," Japan's METI and Ministry of Education, Culture, Sports, Science and Technology wrote in a March report. "METI, which is responsible for information policy, was late to realize the importance of mathematics," the ministries admitted.
The Japanese government plans to reverse the trend, setting a goal in March to foster 250,000 AI experts per year as early as 2025 by requiring nearly all science and some humanities majors to learn related subjects in university.
The lack of diverse talent also forces Japanese companies to recruit from overseas. "Although Japan has a shortage of talent, it is a popular place to work. First-rate people from around the world can be hired with high salaries," said Kenji Nonaka from McKinsey consultancy. "The key will be improving their working conditions, such as creating a forward-looking corporate culture in which AI experts can work."
Japan is aiming to build a free-flow zone for data with the U.S. and Europe. To do so, it will be need to foster more AI specialists and develop an internationally competitive research and development framework in the field. Advanced AI systems' ability to analyze massive amounts of data can also birth to new businesses.