AI in China
AlphaGo, the artificial intelligence program developed by Google, has just won the first of three games against China’s Go prodigy Ke Jie in Hainan. This follows Alpha Go’s defeat of Lee Sedol, another top player of this ancient board game in March 2016. With its complicated positioning and endless possibilities, Go is supposed to be more difficult than chess for AI program to master.
But AlphaGo has other competitors in China. One of them is Jueyi, or FineArt, developed by Chinese tech giant Tencent. After winning 10 straight games against Ke Jie, the world’s No. 1 player, in February, the program won the championship title in the 10th Computer Go UEC Cup held in Japan in March. Although Alpha Go was absent, the event attracted 30 of the world’s best Go AI programs, including Facebook’s Darkforest, Japan’s Deep Zen Go and France’s Crazy Stone.
Jueyi represents the rising ambitions of Chinese tech companies in the emerging field of artificial intelligence, previously dominated by US companies.
China’s AI sector, like most forms of computing in China, is led by the BAT, an acronym referring to China’s Internet triumvirate Baidu, Alibaba Group and Tencent Holdings. Baidu is known for its dominant search engine, Alibaba Group for the country’s top e-commerce platform, and Tencent for social networks and gaming.
With an estimated 730 million Internet users and 630 million smart phone users in China, the development of the Internet in recent years has allowed the three companies access to a huge volume of data, a major advantage in AI research and development.
Baidu was the first Chinese company to open a US-based AI lab, launching its site in Silicon Valley in 2014 with a staff of 200 researchers. It opened a second lab in March 2017, with a staff of 150 people. The company has been focusing on self-driving cars, where it hopes to become a major competitor of Uber and Alphabet (Google’s parent company).
Tencent Holdings established an AI lab in Shenzhen in April last year, which has 50 computer scientists and 200 engineers working on fundamental research and the practical applications of AI. The firm opened its own US-based AI lab in Seattle in May, which will focus on research in speech recognition.
The Alibaba Group’s AI ambitions are built upon its growing cloud service, which grew by 115 percent in 2016. On March 28, Alibaba Cloud, the company’s public cloud division, announced two new AI services, ET Medical Brain and ET Industrial Brain after its flagship AI program ET, along with a new AI platform, called PAI 2.0, which the company said will “allow businesses without an AI background to make practical use of Alibaba Cloud’s ET programs.”
Besides the BAT, other Chinese tech companies have also substantially increased their investment in AI. Lenovo, for example, opened its first AI lab in March. According to CEO Yang Yuanqing, the company will invest US$1.2 billion, or over 20 percent of its total R&D development spending, in the research and development of artificial intelligence, big data and Internet of Things (IoT) technology in the next four years.
In the meantime, the AI sector has become increasingly attractive to investors. According to a report released by Wuzhen Institute, a domestic think tank, the financing received by artificial intelligence companies between 2012 and June 2016 reached US$2.6 billion. Although the figure lags far behind the financing received by American AI firms at US$17.9 billion, it places China well-ahead of other countries, as Britain trails in third place with financing of US$800 million. In terms of AI-related patents, although the US still leads the world with a total of 26,891 patents ahead of China and Japan, which have a total of 15,745 and 14,604 patents, China has surpassed the US in terms of the number of patents newly obtained since 2012.
According to another report released by iiMedia Research, an industry consulting company, the size of China’s AI industry increased by 43.3 percent in 2016 and is projected to increase by 51.2 percent in 2017.
While AI technologies have attracted a large volume of investment from the private sector, the industry has also caught the attention of the Chinese government. In March 2016, the same month Alpha Go defeated Lee, China’s 13th Five-Year Plan (2016–20) called for breakthroughs in artificial intelligence. In May, the National Development and Reform Commission (NDRC) released a three-year action plan on “Internet Plus and AI.”
2016 also saw closer cooperation between local government and AI firms in pushing forward the application of AI technologies in various sub-disciplines, particularly in Zhejiang, China’s prosperous coastal province, where Alibaba Group’s headquarters are based.
For example, last September, the Zhejiang government adopted Alibaba’s speech recognition technologies in all of its 105 courts. Then in November, Alibaba and the municipal government of Hangzhou, capital of Zhejiang Province, launched the initiative “Hangzhou City Brain,” as part of Alibaba’s Smart City project. By applying the company’s big data analytics capabilities and its video and image recognition technologies, a pilot program was launched in the city’s Xiaoshan district last September, which the company said has helped to increase the average traffic speed by 3 to 5 percent.
But for many Chinese entrepreneurs, the support from the government is far from adequate, with many calling for a national strategy and a top-down action plan to support and coordinate the research, development and application of AI technologies. The issue became a major topic during the annual sessions of the National People’s Congress (NPC), China’s parliament held in March, during which the government unveiled its major economic policies.
“AI will be the most important technological revolution in the next ten years, which could reshape the human world,” Lei Jun, chief executive of smart device maker Xiaomi, who also serves as a delegate of China’s National People’s Congress told reporters at an NPC press conference on March 6. Pointing to China’s access to a large volume of data, Lei said China has the potential to become a global leader in AI technology.
But to achieve this goal, Li argued that the government should formulate a national strategy to address various obstacles that cannot be solved by the private sector alone. For example, a major issue for China’s AI firms has been the lack of talent.
It is estimated that there are now only about 30 universities throughout China that provide advanced education on AI, which is far from enough to meet the needs of China’s booming AI sector. It is a major reason why many Chinese AI firms have chosen to open labs in the US.
Lei argued that the government should adopt policies to devote more resources to develop talent at home and attract talent from abroad, along with policies to support the application of AI technologies.
Lei’s call was echoed by Liu Qingfeng, the chairman of iFlyTek, a company specializing in voice recognition, another NPC delegate, who argued that the government should take a strategic approach to the AI sector with policies fostering better collaboration among enterprises to promote the development of the entire AI ecosystem in China.
It is clear that the voices of entrepreneurs did not go unheeded. In delivering the annual government work report on the NPC, Chinese Premier Li Keqiang said that the government will “make plans for the development of strategic and new industries” including AI and other industries. It is the first time that AI has made its way into the government’s annual work report, and many believe that a national AI strategy is now gradually taking shape.
On March 2, just a few days prior to Li’s speech, China unveiled its first national AI lab, which focuses on deep learning. Financed by the NDRC, the lab will be operated by Baidu in cooperation with Tsinghua University and the Chinese Academy of Sciences. In January this year, Alibaba and the National Astronomical Observatory of China had also jointly set up a data and research center for astronomy.
On May 13, a national laboratory for “brain-like AI” technology was inaugurated in Hefei, capital of east China’s Anhui Province. Approved by the NDRC, the lab is based at the China University of Science and Technology and will be operated in collaboration with other top AI research bodies such as Fudan University, Shenyang Institute of Automation of the Chinese Academy of Sciences, as well as Baidu.
On April 25, the NDRC, Ministry of Science and Technology and the Ministry of Industry and Information Technology jointly issued a mid- and long-term development plan for the automobile industry, setting a goal of having 80 percent of new automobiles equipped with driving assistance, or partial and conditional automatic driving system by 2025.
The vision appears to coincide with that of Baidu. Only a week earlier on April 18, Baidu announced its Project Apollo, an autonomous driving platform that will include both software and hardware, which the company aims for use on urban roads towards the end of the year and on highways and open city roads by 2020.
Analysts believe that these developments indicate the prospects of greater cooperation between the government, universities, national research institutes and AI companies in the future.
It is also reported that following the Chinese Academy of Engineering’s proposal of the “Artificial Intelligence 2.0 Plan,” the Ministry of Science and Technology has been working to draft a plan for the development of artificial intelligence through 2030, which could be released within the year.
With China’s ambition to build itself into the world’s leading AI powerhouse, the global competition in the futuristic industry may just be starting.