The State Council (the cabinet) has unveiled a development strategy designed to make China the world's leading artificial intelligence (AI) power by 2030. China has several advantages, including access to reams of personal data, a developed server infrastructure and improving talent pool, to create an economy powered by AI. But risks remain, especially relating to the hype that is driving investment in the sector. It is also unclear whether a top-down development plan will work in an industry that has been mostly pushed forwards by private companies in recent years.
The State Council plan, unveiled in July 2017, aims to expand the annual revenue generated by the AI industry by seven-fold to reach Rmb1trn (US$148bn) by 2030. It laid out tax incentives and research and development (R&D) cost deductions for AI small and medium-sized enterprises and start-ups. The plan also committed to strengthening the protection of AI intellectual property and developing a public patent pool. Moreover, the government will create industrial parks across the country to nurture AI companies, and AI innovation application pilot zones to support experimentation. Companies in all sectors will be pushed to integrate AI technologies into their operations.
China's advantages in AI
China has quickly built up its AI capacity. Annual AI-related patent applications in China ran to over 8,000 between 2010 and 2014, triple the level of the previous five years. China is catching up to the US in AI. When setting out its own AI development plan in October 2016, the US government noted that China had overtaken America in the number of published journal articles on deep learning, a technique for implementing machine learning. According to Wuzhen Institute, a Chinese technology-focused think-tank, Chinese AI firms received US$2.6bn in funding in 2012–16. Although less than the US$17.9bn poured into their US peers over the same period, the total has been growing quickly.
In some senses, China has an ideal environment for developing AI. The country has 730m smartphone users, more than any other country in the world. Chinese consumers are accustomed to paying for goods and services online, using digital assistants and voice-recognition software, and do not seem overly concerned about privacy or data sharing. No other country has generated such quantities of data that can facilitate machine learning and make AI-related technologies possible. The government's development plan plays to these strengths, noting plans to explore how to unlock the commercial value of public data.
The largest private domestic technology firms are already pushing into AI, seeing enormous potential in the field. The so-called BAT—the big three Chinese internet and e-commerce companies, Baidu, Alibaba and Tencent—are investing heavily in AI, from facial recognition to messaging bots. Baidu, China's leading search engine, announced in September 2017 an Rmb10bn (US$1.5bn) "Apollo fund" focused on backing autonomous-driving technology companies. The fund is part of the company's plan to compete with US rivals, such as Tesla and Waymo, the self-driving car unit owned by Google's parent company, Alphabet.
Both Baidu and Tencent have set up their own AI research labs in China and the US. Although unclear how far the development has gone, Tencent has also started developing its own autonomous-driving system, in an attempt to use its mapping and AI technology to compete with rivals Baidu and Google. In October Alibaba announced the launch of DAMO Academy, a programme that will set up seven R&D labs worldwide researching AI, quantum computing and more.
Emerging talent pool
Yet data alone will not be enough for China's AI ambitions. Although it has an abundance of computer programmers and a historically high focus on maths education, China has a talent shortage when it comes to top-tier AI experts. As a result, the largest domestic companies have been aggressively tapping the Silicon Valley talent pool. Baidu hired Andrew Ng, Google Brain's founder and a leading AI expert, as its chief scientist in 2014. After he left Baidu in 2017 to launch a new business, Baidu hired Lu Qi from Microsoft. Mr Lu is leading a team to build "Baidu Brain"—a system which offers 60 different types of AI services, as well as an AI ecosystem that would enable other AI companies and developers to accelerate their pace of innovation.
Tapping the global talent pool can help to plug gaps in the short term, but the longer-term development of China's AI sector will depend on cultivating local resources. Encouragingly, some major universities have already launched AI programmes, such as Peking University, Tsinghua University and Zhejiang University. The State Council plan openly calls on universities and vocational schools to offer AI skills training programmes. It also says that AI-related courses will also be added into the curriculum of primary and secondary schools.
Venture capitalists are pushing a lot of cash into AI companies. According to KPMG, in 2016 China saw venture-capital investments surge to a record high of US$31bn, marking 19% growth. Many provincial governments in China have also set up funds, called "government guidance funds", for investment in local start-ups. For example, in 2017 the Hubei and Jiangsu provincial governments set up Rmb40bn and Rmb80bn industrial funds for start-ups, respectively.
But investors should be wary. The market potential for AI is considerable, but there is a lot of hype and a lack of substance behind many AI start-ups. AI is typically based on complex models and technology, such as natural language processing and deep learning, which take years to develop. Many Chinese start-ups claiming to use AI in fact do not use it.
The development of AI in China has been mostly led by large private internet companies, principally the so-called BAT. But the State Council development plan shows that the government intends to strengthen its role in allocating resources in the industry. China's state venture-capital funds are already surging. According to Sun Hung Kai Financial, a brokerage and advisory firm, government guidance funds have already surpassed private venture-capital funds: as at September 2016 government guidance funds had raised a total of Rmb3.3trn, compared with Rmb2.2tn raised by private funds. Whether the government can invest as efficiently as the private sector remains to be seen, but on past evidence from other sectors it will not.
China also lags behind in the moral discussion of developing AI. Despite all its potential benefits, the use of AI raises questions about data privacy and the potentially far-reaching application of its technology (autonomous weaponry is one focus of AI development). On these points, the public discussion in China is not yet as advanced as in other countries.