UN Study: China, US, Japan Lead World AI Development

A United Nations agency says the United States, China and Japan are leading the world in developing artificial intelligence, or AI, technology.

The three countries were named in a new study from the UN’s World Intellectual Property Organization, or WIPO. The organization says on its website that WIPO aims to “lead the development of a balanced and effective international intellectual property system.”

The new study examines development activities for different AI technologies worldwide. It also attempts to predict the future progression of AI as it relates to new inventions.

The study is based on international patent requests, legal findings, scientific publications and other information. It includes patent requests in machine learning through 2016, the last year for which details are available. Much of today’s AI-related technologies are powered by machine learning.

The study found that there had been as many patent applications for artificial intelligence since 2013 as there were in all the time since the term AI was first used in the 1950s.

Machine learning was the most common method of AI listed in patent requests. It was responsible for more than one-third of all identified inventions.

Machine learning uses computer algorithms to teach machines to make independent decisions by learning from earlier or existing behavior.

Patent requests for machine learning activities grew on average by 28 percent a year between 2013 and 2016, the study found. It also showed that 26 of the top 30 AI patent requests came from businesses. Universities or public research organizations made up the rest.

The study found that U.S. technology company IBM had by far the largest number of AI-related patents, with 8,920. Next came U.S.-based Microsoft Corporation, with over 5,900.

At number three was Japan’s Toshiba Corporation. Next came South Korea-based Samsung Group and Japan’s NEC Group.

WIPO reported that China had 17 of the top 20 academic organizations filing for AI-related patents. It noted 1 China was especially strong in the fast-growing area of “deep learning.” This is a machine learning method that includes speech and facial recognition 2 systems.

The report said that deep learning methods increased from just 118 patent requests in 2013 to nearly 2,400 in 2016.

Another kind of machine learning, called neural 3 networks, is also “revolutionizing” AI, the report said. A neural network is a computer system built to simulate 4, or act like, the human brain. Patent filings for neural networks grew at a rate of 46 percent from 2013 to 2016, reaching 6,506, the study found.

The single most popular AI technology identified in the report was computer vision. It was listed in 49 percent of all AI-related patents, the study found. Computer vision is used in image recognition systems and also in self-driving technologies.

Francis Gurry is director of WIPO. He said the study provides clear evidence that AI technologies are growing at a faster rate than ever and will continue to do so.

The result, Gurry said, will be many new AI-based methods and products that can change lives across the world. Such technologies will also “shape future human interaction with the machines we created,” he added.

I’m Kelly Jean Kelly.

Words in This Story

intellectual property – n. an invention, such as a written work, design or piece of music, that is a product of creativity

artificial intelligence – n. ability of a machine to use and study computer data in an attempt to reproduce 5 human behavior

patent – n. a legal right a person or company receives to make or sell a particular product so others cannot copy it

algorithm – n. set of steps that are followed in order to solve a mathematical problem or to complete a computer process