China and its Ambitions in Artificial Intelligence
In the summer of 1956, four American scientists, John McCarthy, Marvin Minsky, Nathaniel Rochester and Clause Shannon, met to give a lecture at Dartmouth College. The term Artificial Intelligence was used for the first time in the history of mankind (CEA & Esprit Sorcier 2018).
The scientist Marvin Minsky gave a definition of artificial intelligence that I find simple but very precise: Artificial Intelligence is “the science of making machines do things that would require intelligence if done by men (Minsky 1956). The idea is to create a machine, that will be able to think as we do, and live as we do.
How it works
The concept of artificial intelligence is something that speaks to everyone. Especially because it is very present in the science fiction works of art that we all love. Who does not know HAL9000, this computer, equipped with the best artificial intelligence ever created by man, which embodies by himself the hopes and fears that artificial intelligence evokes for humans (Kubrick 1968).
Artificial intelligence has also been a high-profile subject for several years, newspapers have devoted articles to it, television channels have made it into programmes and all major high-tech companies have projects related to this new technology. New technology?
Artificial intelligence does not date back to our current millennium, more than two thousand years ago, Aristotle already presented the idea that instruments can perform various tasks alone to free human time (Aristote s.d.). After that, many philosophers and scientists added new theories and thus clarified the concept of artificial intelligence. However, all accelerated during the 20th century, especially with Alan Turing. Turing is an English mathematician who achieved many achievements during the Second World Warn such as deciphering the Enigma Code, and who wrote a paper called “Computing machinery and intelligence” in 1950 (Berny 2017). In this paper, Turing proposes a test: the “imitation game”. The purpose of this test is to test a machine’s ability to simulate a human conversation, thus introducing to the world to the idea that a machine would one day be able to think, like human.
As we saw, artificial intelligence has been highly publicized in recent years, thanks in particular to a few achievements that have caught our mind. In 1997, DeepBlue, developed by IBM, defeated world chess champion Garry Kasparov. In 2016, AlphaGo (Google) achieved another accomplishment, which remains for a long time impossible, by defeating a champion of the game of Go: Lee Sedol. (Clubic 2017)Such achievements like this are many, but often the companies that made his achievements possible are Western.
However, China also has assets and desires in the field of artificial intelligence and this for the next few years, although it is less known. Then we can ask ourselves the following question:
What are China’s ambitions in terms of artificial intelligence by 2030?
State of the art
Today, artificial intelligence can be found in many fields. Whether it is in autonomous cars that are becoming more democratic, in the facial recognition of our phones, when we do research on the Internet and even in the robots that our industries use.
These technologies, even if they are not always true artificial intelligence (in the sense of Turing), they are taking up more and more space in our lives.
They use concepts such as Deep Learning (technologies that use a large amount of data to adapt their reactions without human intervention), and are developed by private companies. More than 3,000 start-ups worldwide are currently working in these technologies. The United States owns a large number of these companies, China is in second place. However, of the 10 largest companies in this sector (by value) five are Chinese! The first, which worth $4.5 billion, is also Chinese (Robles 2018).
Although we mainly hear about GAFAM (Google, Apple, Facebook, Amazon and Microsoft) in the newspapers (all American’s companies), the truth is that China has a strong potential for artificial intelligence through many private companies. And we will see later that the Chinese government is supporting these companies financially (Lee 2018).
One thing is certain, in terms of artificial intelligence the United States and China dominates the sector of research: in 2015, these two countries had more than 10,000 research papers, while for the other countries the figure is 5,000. The match is between the USA and China.
Let us take a closer look at these two countries by putting them in parallel. As we have seen, the United States has the most companies specialized in artificial intelligence, China is second. However, in terms of the number of research papers written, China is ahead of the United States (Robles 2018). But even if China writes more research paper, it does not mean that they have more influence. Indeed, if we look at the graph below, we see that the level of influence of publications (H-index??) is 373 while it is only 168 for China. This means, for example, that Chinese publications are less cited than American ones. How can we explain this result? When we look at the artificial intelligence talents (understand a specialized searcher in artificial intelligence), we observe that China has about 8.9 per cent of those talents while United States have 13.4 per cent. China also don’t have the “top talent” that United States have (Dai et Shen, ‘Made in China 2025’: China has a competitive AI game plan but success will need cooperation 2018). This is the main issue for China today.
It is therefore very difficult to say which of these countries has the most influence, they may be too close to each other to conclude.
Other indicators allow us to analyse the situation. The country with the most patents on artificial intelligence is China, just ahead of the United States. Between 2017 and 2018, China’s artificial intelligence industry experienced a 67 percent increase in the production of patents and research paper.
China also dominates the sector in terms of financing new start-ups. 48 per cent of new companies are in China, when 38 per cent in the United States.
However, China suffers from a second major flaw (the first being the lack of talent). China does not have a complete ecosystem in the artificial intelligence industry. When we look at the graph below, we see that the major Chinese start-ups are only present in three different sectors of activity: health, sales/commerce and communication algorithms. While the United States has companies in almost every field related to IT: from finance to robotics to the Internet of Thing (Barton, et al. 2018).
This type of ecosystem is very advantageous for countries, it creates an environment conducive to research and innovation by facilitating the latter. While the United States benefits from such an environment, it will be difficult for China to re-create a similar one (Barton, et al. 2018).
When we look at all these figures, we see that China is making extremely rapid progress in the field of artificial intelligence. However, all this is very recent and China is lagging behind the United States.
But how long will this delay last?
China has many assets: a very large population, more than one billion and four hundred million inhabitants (Country Meters 2018). This offers Chinese companies considerable amounts of data because privacy is not a priority. However, as we have seen, having data is essential in the field of artificial intelligence, Deep Learning bases its operation on the analysis of a large amount of data.
In addition, Chinese companies have long been considered as companies that copied. However, many Chinese companies have been able to stand out and become real giants in the industry, such as Alibaba, which has been able to innovate to adapt to the Chinese market.
Another of China’s assets is its government. Indeed, the Chinese government has many ambitions in the field of intelligence and relies heavily on it. As we will see later in this document: China plans to become the world leader in artificial intelligence by 2030 (Wenderoth 2018).
China’s Artificial Intelligence plan
On July 20, 2017, the State Council of China presented “the next generation artificial intelligence development plan” (He 2017). This plan sets out different objectives for China. The final goal is clear and there is no lack of ambition: to become the undisputed world leader in the field of intelligence by 2030.
For the Chinese government, artificial intelligence “will profoundly change human society and life and change the world” (State Council of China 2017). The objective of this plan is therefore to bring China into this new world to benefit from its advantages.
What are the advantages that China is looking for? Why China plan to become a leader in artificial intelligence?
Several answers can be provided, and this in three distinct fields: economy, military and social.
First of all, China could benefit from a very positive economic fallout for their economy. According to a McKinsey Global Institute study, at least “half of China’s work activities could be automated” (Barton, et al. 2018), making China the country with the greatest potential for automation in the world (Barton, et al. 2018). This potential automation can boost China’s productivity. Especially if you considered the current demographic trend of China: there is not enough birth to face the ageing of the population (Leptre 2016). Automation may counterbalance the economic impact of this problem. Moreover, PWC’s research during the year 2017, shows that artificial intelligence may boost the GDP if China by 26 per cent (Ding 2018)!
If there is one area in which artificial intelligence can be frightening, it is the military field. Indeed, artificial intelligence promises a technological revolution that will be a defence challenge. Several applications are being considered in the short term by armies around the world. Artificial intelligence will make it possible to manage and simulate an environment, to detect, simplify, automate and even predict (NOEL 2018)! Autonomous fighting machine are already developed in the world and many weapons companies integrate artificial intelligence in their product. So artificial intelligence represents for China a way to catch up with the United States, to improve its defence capability, … This is a strategic challenge for the military field.
The third aspect for which China is seeking to become a leader is the social aspect. The Republic of China is not recognized in the world as a republic that respects human rights: freedom of expression and information does not exist, everything is controlled by the Communist Party, there are many “suspicious disappearances” and China is often accused of having “re-education” camps (Human Rights Watch 2018). Artificial intelligence could become a tremendous asset for China in this context. This would allow China to implement new, increasingly automated and efficient surveillance systems (Abbany 2018). Surveillance would be more advanced than ever.
This surveillance and new espionage techniques can also play an important role in a military context.
As we saw, there are many reasons why China wants to develop its artificial intelligence industry and those reasons will offer China serious economic, social and military advantages. However, the government’s ambitions are important, especially since the period to achieve them is quite short (only 19 years).
How does China intend to achieve its objectives and reach its ambitions for 2030?
In order to achieve its ambitions, the government has implemented various action within their “next generation artificial intelligence development plan”.
The first of these strategies is to build a legal and ethical environment around the development of artificial intelligence. This development cannot be achieved without a legal framework: The United States is ahead of the curve with many autonomous cars circulating on the roads of different states. China therefore wants to set up a legal framework to allow these companies to develop new technologies. The Chinese government is also seeking to establish standards on the moral choices that artificial intelligence may have to make: should an autonomous car run over an elderly person or a twenty-year-old person? (Barton, et al. 2018)
Secondly, policies in favour of companies in this sector will be put in place to facilitate their development, including economic policies such as tax cuts (He 2017).
As we have seen, China suffers from two major problems that undermine its objective of becoming the leader: the lack of talent and the absence of a complete eco-system. Two strategies will then be implemented.
The first is to address the problem of lack of talent. To this end, China will improve its artificial intelligence education system and facilitate the integration of these new talents into the world of work. To this end, the Chinese government will encourage and support universities, companies and other research centres (State Council of China 2017).
A second strategy aims to develop an eco-system. To do this, China wants to set standards, share public data to encourage their companies to cooperate. The aim is also to make it easier for Chinese companies to collaborate with foreign companies (Dai et Shen, ‘Made in China 2025’: China has a competitive AI game plan but success will need cooperation 2018)
The latest strategy implemented is to facilitate the integration of technologies related to artificial intelligence into China’s “classic” industry. This is intended to bring positive economic benefits more quickly and thus encourage companies to invest more in artificial intelligence while developing the country’s industry.
To carry out these various actions, the Chinese government is investing significant sums in its companies. Indeed, China plans to invest $29.8 billion by 2020 in research. This figure is expected to reach $79.48 billion by 2030 (Zhang 2017).
In addition, China is also investing to ensure that partnerships between private companies and the public sector can take place (Abbany 2018).
Regarding the lack of talent, China relies on its private companies to help it by facilitating the recruitment of new talent. Thus, students can be paid $115k when they work in the artificial intelligence sector after their studies (Becoming Human 2018).
Although China has integrated an ethical dimension into its “next generation artificial intelligence development plan”, it has not integrated any measures to protect the data of Chinese citizens as the European Union and its DGPS can do. This makes it easy for companies to use user data. And as we have seen, data are essential elements for artificial intelligence (State Council of China 2017).
The Chinese government will also better defend its companies when it comes to intellectual property (patents) (State Council of China 2017).
As we have seen, China is already a leader in the field of artificial intelligence. Second after the United States but with much higher growth in this sector. However, it does not intend to stop and present a very ambitious plan to become the world leader by 2030.
This is an almost impossible challenge for any other country to meet, but I think China has the means to do so. Although it has lagged considerably behind the United States for decades, this lag tends to disappear.
Especially since China is giving itself the means to succeed, especially with colossal investments, much more than what France is planning: 25 million euros shared between five and ten companies (Olivier 2017).
The Donald Trump government does not seem to consider artificial intelligence in the same way as China. Indeed, the United States provides very little funding for companies and leaves the private sector to manage itself alone.
However, China must be careful, the development of these new technologies risks to disrupt China’s current model very (too?) quickly. As we have seen previously, a large part of Chinese work can be automated: there is therefore a real risk that the number of jobs requiring few qualifications will fall much faster than the Chinese population is aging. This could increase inequalities on Chinese territory, inequalities that are already well established….
Artificial intelligence in the military field is an interesting subject to focus on. These technologies can breathe new life into weapons and make our current weapons much more dangerous than they are today. I believe that an arms race is in no one’s interest and that cooperation between different nations in the development of artificial intelligence is the best way to minimise the risks.
It is also important that each country, each company works for the common good and that artificial intelligence does not become the “worst event in the history of our civilization” as Stephen Hawking has said.
To give a final answer: China definitively want to become the undisputed leader before 2030 and they have, during the moment I wrote those lines, the capabilities do achieve their goal.