In 2016, artificial intelligence defeated a professional player at Go for the first time. The AlphaGo program, developed by DeepMind, first beat three-time European Champion Fan Hui, and a short while later bested Korean Lee Sedol, considered to be one of the best Go players in the world. Many believe that it was artificial intelligence’s conquest of the ancient Chinese game that convinced Chinese authorities to think seriously about including the development of artificial intelligence among its strategic goals. However, the importance of artificial intelligence is not a new idea. Although the foundational Next Generation Artificial Intelligence Development Plan came out in June 2017 a year after the triumph of AlphaGo, the field had already been earmarked as a priority in many earlier documents.
21st Century Electricity
Back in 2006, the State Council of the People’s Republic of China published its National Medium- and Long-Term Program for Science and Technology Development for the period up to 2020. It names smart sensors, intelligent robots, and technologies of augmented reality as areas of priority. The Made in China 2025 plan was released in 2015. It includes a USD 300 billion fund for the development of high technology and industrial manufacturing. The plan involves intensifying work in R&D, new materials, artificial intelligence, the creation of fifth generation telecommunications networks, and the manufacturing of robots. In 2016, the State Council published its Guiding Opinions on Actively Promoting the Internet Plus Action Plan. The document prioritizes artificial intelligence, along with big data, blockchain, and machine learning, for a state strategy aimed at accelerating the use of information and communication technologies for the development of the smart industry.
Why has China become so serious about developing artificial intelligence? Artificial intelligence has been called the electricity of the 21st century. It is a technology capable of spanning numerous sectors and giving rise to a new industrial revolution, which is exactly what China needs now. For some time, the country has functioned as the world’s factory; by employing cheap labour and copying foreign technology, China has supplied the entire world with inexpensive (and not always high-quality) manufactured goods. It is a model that has facilitated several decades of double-digit economic growth rates.
However, China has now fallen into the trap of average income. The population’s living standards have grown, reducing the country’s main competitive advantage – cheap labour – to naught. As a result, the only solution is to compete with developed economies and with their qualified personnel and innovations. To do so, the country needs a technological breakthrough. Artificial intelligence could play just such a part.
Russian President Vladimir Putin has noted that the leader of artificial intelligence will be the ruler of the world. Chinese authorities understand this. In June 2017, the Chinese State Council released its Next Generation Artificial Intelligence Development Plan, which states that artificial intelligence has become a new arena for international competition. It is a strategic technology that will set the stage for future development and determine a country’s international competitiveness, national security, and influence in the world. Consequently, the Chinese authorities hope artificial intelligence will help transform their economic growth model and expand geopolitical influence, as well as modernize the army and strengthen the country’s defence capabilities. Chinese President Xi Jinping has repeated on more than one occasion the importance of civil-military integration and the need to remove barriers between the commercial economy and the military-industrial base. In other words, artificial intelligence is seen as a dual-use industry. It seems that the development of military and civilian uses for these technologies will occur simultaneously.
The Chinese Next Generation Artificial Intelligence Development Plan sets three strategic goals. The first is for Chinese AI to match pace with corresponding sectors in key developed countries by 2020, with the foundational branch of AI collecting USD 22.5 billion and related industries exceeding USD 150 billion. The second goal is for China to take the lead in some areas of AI by 2025, with the foundational branch collecting USD 60 billion and those related attaining USD 745 billion. Finally, by 2030, China should become the world’s main centre for innovation in the field of AI, with the foundational branch collecting USD 150 billion and related branches reaching USD 1.5 trillion.
The programme does not explain exactly how to achieve these strategic goals, although it does officially have a section devoted to the topic, titled “objective tasks.” It is essentially a list of industries amenable to the introduction and development of artificial intelligence. These include smart cities, AI in medicine, swarm intelligence and deep semantic analysis, computer vision, and the use of AI in the defence industry and social management. It seems that the programme is not so much a practical guide to action as a general reference point for central and local authorities. Officials can select the area of the programme best suited to their particular region and develop it by adding their own initiatives. For example, authorities in the historically poor province of Guizhou chose to make it China’s data centre, enabled by the favourable cold climate and mountainous terrain. Chinese technology giant Tencent is already building a giant 30 sq. km underground data storage in the mountains of Guizhou. A similar data centre is being built in Guizhou for Apple’s iCloud. Construction is to be completed by 2020, and the centre will occupy an area of 67 hectares. Investments in the IT sector in Guizhou have grown by 378 percent. USD 2.8 billion was invested in Guizhou last year in services related to the transfer, storage, and processing of data. What was once the poorest province has become one of the few in the first quarter of 2018 with double-digit GDP growth rates of 10.1%.
The Next Generation AI Three-Year Development Plan released by the Ministry of Industry and Information Technology of the People’s Republic of China in December 2017 is more specific about how to develop artificial intelligence and covers a period extending up to 2020. It has set several tasks. The first is to stimulate the development of smart products. This includes, in particular, cars connected to the Internet, smart robots and drones, voice and facial recognition, and machine analysis of medical images. It also calls for a breakthrough in key fundamental technologies like the development of chips and neural networks and open source platforms. In addition, the plan provides for the development of industries using key technologies from artificial intelligence.
The three-year plan coincides with another document issued by the Ministry of Science and Technology of China listing 13 technology projects with priority over large-scale public investment. These projects are to be implemented before 2021. The most notable among them calls for the creation of an artificial intelligence chip that promises to be 20 times more productive and energy efficient than the American-produced NVIDIA Tesla M40 – one of the most widely used artificial intelligence chips at the moment.
Where Can I Get Five Million Scientists?
It’s no wonder that the Chinese authorities are focusing on chips in the development of artificial intelligence. They pose the greatest problem at present. The development of microchips is an extremely knowledge-intensive and capital-intensive process, and the results are not always manifest. China is extremely dependent on imports of foreign microchips, mainly American ones, with only 16% of its chips produced in China itself. Annual imports amount to USD 200 billion, which exceeds Chinese oil imports. China’s share in the world market is even smaller: in 2015, China accounted for only 4% of world chip production, while the USA accounted for 50%. Naturally, the Chinese authorities support local AI manufacturers and developers in every way possible, granting them tax breaks, administrative preferences, and direct financing. For example, the China Integrated Circuit Industry Investment Fund collected USD 31.5 billion. This is not, however, such a large amount in the industry. Intel alone spent USD 12.7 billion on research and development in 2016.
China is also experiencing an acute shortage of qualified personnel. According to estimates from the Ministry of Industry and Information Technology of China, about 5 million specialists will be needed to implement the tasks that have been set. At the present time, there are 1.9 million professionals specializing in the field, 850,000 of which are in the USA, with only 50,000 in China. More than 43% of those in China came from the USA. Approximately 2,500 companies worldwide are engaged in the research, development, and practical application of artificial intelligence. The Tencent Research Institute has acknowledged that US companies occupy the lion’s share of the market and outstrip China in all aspects of AI research and development. Thirty-three American and fourteen Chinese companies are occupied with the development of processors and chips. Of the companies involved in natural language processing, computer vision, and image recognition, 586 are American and 273 are Chinese. And finally, 488 American and 304 Chinese companies are working on machine learning, smart drones and robots, and self-guided cars.
The only area in which China enjoys an undeniable competitive advantage is in the colossal amount of data it possesses. The population of China is considerable, and more than half of it – 752 million people (twice the population of the United States) – uses mobile Internet. Eighty-four percent regularly make use of mobile payments. These people leave “digital footprints” behind them in their everyday lives. This is precisely the kind of big data that is so important for machine learning.
The Rest of the World Will Help
In all other areas, China relies on foreign technology and personnel. The Next Generation Artificial Intelligence Development Plan and the Three-Year Action Plan discuss the need to encourage Chinese companies to carry out mergers and acquisitions of foreign partners. This policy has been successful. A few years ago, Chinese tech company Baidu opened its Silicon Valley Artificial Intelligence Lab, and in 2017, the company opened a second centre there for research and development of self-guided cars as part of the Apollo project. Soon after that, the company’s third centre in the US, the Business Intelligence Lab, opened for research and development of big data. The other tech giant, Tencent, opened its artificial intelligence research centre in Seattle.
Foreign companies are happy to open research laboratories in China itself. Google is hiring employees for its research centre in Beijing, even though the company’s main products, the Google search engine and Gmail, are blocked in China. On the other hand, Chinese authorities are trying to create favourable conditions for the work of foreign experts. Scientists and developers in the field of high technology can obtain a Chinese visa for a period of 5 to 10 years with the ability to enter the country an unlimited number of times. Moreover, Chinese companies spare no expense to attract specialists from foreign companies and competitors: a high-level scientist in China can receive up to a million dollars a year.
The US is concerned that China will borrow American high technology and attract US scientists to work in China, which will eventually lead to Chinese supremacy in the field of AI. In the US, the development of artificial intelligence is carried out mainly by private companies, and these companies often do not agree on how to develop dual-purpose solutions and products. For example, when Google took over DeepMind, the latter forbade the use of their products in the military or to monitor citizens of the country. Moreover, when Google acquired the robot developer SCHAFT, the company declared that it would not work for the Defense Advanced Research Projects Agency.
The situation in China is completely different from that of the US. Despite the fact that a third of the world’s tech startups with capital exceeding USD 1 billion are present in China, three technological giants dominate all the rest: Baidu, Alibaba and Tencent (BAT). In most of the start-ups, direct or indirect investments have come from BAT. China’s Ministry of Science and Technology formed the first working group for the development of next-generation artificial intelligence by these companies. In the group, Baidu will be responsible for self-guiding cars, Alibaba for smart cities and city think tanks, and Tencent for computer vision. What’s more, BAT has made no bones about sharing big data with the state if necessary and opening party committees within the company itself. The possibility of formalizing these relations by means of the government’s acquisition of a 1% stake in the companies has also been discussed. When Chinese President Xi Jinping speaks about the need for civil-military integration, it can be assumed that all the achievements of Chinese (and foreign) scientists and companies in the field of artificial intelligence will become available to the military.
The Race for Artificial Intelligence
There are no programmes directly involved in developing artificial intelligence in the Chinese military. However, according to Elsa B. Kania, an Adjunct Fellow with the Technology and National Security Program at the Center for a New American Security, the Chinese military understands the need for “intelligentization” of the military-industrial complex. Future military actions are likely to be impersonal, intangible, and inaudible. It is her opinion that China is actively developing UAVs, underwater drones, and self-guided combat vehicles.
This has the USA on edge. If the second half of the 20th century saw two superpowers, the USSR and the USA, racing after nuclear supremacy, then the 21st century will see two superpowers, this time the USA and China, racing after artificial intelligence. The US is trying to resist: President Trump has initiated an investigation into violations of intellectual property rights by China under article 301 of the 1974 US Merchant Act. This investigation has shown that China infringed on four main aspects of American intellectual property rights: compulsory transfer of technology, discriminatory licensing rules, cross-border takeovers, and theft of intellectual property. In regards to the “301 investigation”, US Trade Representative Robert Lighthizer said: “These are things that China listed and said we’re going to take technology, spend a couple hundred billion dollars and dominate the world. These are things that if China dominates the world, it’s bad for America.” As a result, Trump announced the possibility of introducing tariffs on goods from China to the tune of USD 150 billion in order to contain the development of the Made in China 2025 programme.
It is true that China has little difficulty parrying these attacks, promising in return to limit imports of American-produced soybeans and sorghum, more than half of which is exported to China. This would seriously impact the American farmers who made up a significant part of the electorate Trump relied on in his election campaign. This has made American attempts at containment thus far unsuccessful. Following the latest round of trade negotiations between the Chinese delegation headed by Vice Premier of the State Council of the People’s Republic of China Liu He, President Trump announced that the introduction of tariffs on Chinese goods has been postponed, as China agreed to lift restrictions on imports of American agricultural products.
China has publicly stated from the very beginning that if it is still possible to work on reducing the American trade balance deficit, then industrial and technological policy and development are an internal Chinese matter not up for discussion. In the eyes of the world community, President Trump has not appeared to be a crusader for justice, but rather an aggressor, encroaching on the basic principles of free trade and the international division of labour. If American companies willingly accept mergers and acquisitions from Chinese partners, then it must be economically profitable for them to do so.
Perhaps in the race for artificial intelligence it would be better to change tactics and move from deterrence to competition? The Obama administration developed an artificial intelligence programme and, as Western media outlets have noted, the Chinese programme for the development of next-generation artificial intelligence, established just a year after the American programme, appeared in many respects to copy it. In particular, the American programme suggested an increase in public funding for research and development in artificial intelligence. However, the Trump administration has decided to reduce the National Science Foundation’s already trifling budget for research on so-called intelligent systems by 10% to 175 million dollars. Instead of increasing their own spending on research, the US is trying to limit China’s development, but China is unlikely to make concessions. One recent article in a leading Chinese newspaper, the Guangming Daily, urged others not to miss the opportunities of a new technological revolution. It noted that China had been a strong agrarian country but had missed the opportunities provided by the industrial revolution and as a result had become passive and subject to infringements on the world stage. Thanks to the efforts of the last several generations, according to the newspaper, China has come closer than ever to bringing about the rebirth of the great Chinese nation and has never been as sure of itself as it is now. It would appear that the country’s leadership is trying to heed the lessons of the past so as not to miss the historic chance of leading the digital revolution.
First published in our partner RIAC