U.S.-Asia Business
AI in China: Beyond the Hype
By Daniel Allen
Striving to meet hugely ambitious AI targets over the next decade, China is leveraging its strengths and addressing its weaknesses.
Today, artificial intelligence (AI) is widely viewed as a key driver of future growth, competitiveness and job creation, boosting productivity and efficiency and reducing costs. Powered by AI, the world now stands on the cusp of a fourth industrial revolution.
With an ever-increasing number of businesses now employing cognitive technologies, experts predict that by 2030 AI could be contributing as much as $15.7 trillion to the global economy, while a recent report from Accenture suggests that AI could add 1.6 percentage points to the growth rate of the Chinese economy by 2035. It's little surprise, then, that China is vying hard for AI supremacy.
The development of China's AI industry became a national priority in 2017, with the publication of the country's "New Generation Artificial Intelligence Development Plan." According to the document, China's AI industry should be on a par with the AI industries of other leading countries by 2020, pushing the boundaries of research, attracting top talent, and generating RMB 150 billion ($21.5 billion, as of February 2020) from its core operations. By 2030, China aims to be the world's preeminent AI power, with an AI industry worth RMB 1 trillion ($150 billion).
Closing the gap
Whether China can realize its supremely ambitious AI targets remains to be seen. The country's whirlwind development as an AI powerhouse is frequently exaggerated, stirring up emotions that obfuscate the reality of progress on the ground.
In a 2018 study on the rise of AI in China, Jeffrey Ding, a researcher at the University of Oxford's Future of Humanity Institute, awarded the U.S. 33 points and China 17 points (with 100 points representing overall global AI capacity) in a ranking of AI capability. From a multinational perspective, this puts China in second place, but by some margin.
"The hype over China as an AI superpower poised to overtake the U.S. is certainly overblown," says Ding. "The U.S. maintains structural advantages in the quality of scientific and technological inputs, the fundamental layers of the AI value chain, and key subdomains of AI (such as computer vision, predictive intelligence and natural language processing)."
Allison Malmsten, a marketing strategy analyst at China-based consultancy Daxue Consulting, adds, "I'd say China has overtaken the EU, but certainly remains behind the U.S. in the race to develop AI. To continue its upward trajectory, China will need to play to its advantages of having a massive data pool and widespread adoption."
"I'd say China has overtaken the EU, but certainly remains behind the U.S. in the race to develop AI. To continue its upward trajectory, China will need to play to its advantages of having a massive data pool and widespread adoption."
Funding boost
According to a 2019 white paper from global market intelligence firm IDC, China's AI market will be worth around $12 billion dollars by 2023. But while the country may be coming up short on its own development plan targets, it would be extremely short-sighted to dismiss Chinese AI efforts out of hand.
Since the 2017 plan was published, comprehensive government support, coupled with a favorable investment climate and a rising level of talent retention, has seen China's AI industry blossom, and the quality of Chinese AI research has also risen significantly.
"Data, policies and capital are China's major, major strengths in AI," says Taylor Lam, Deloitte China's Technology, Media & Telecommunications industry leader. "With 1.4 billion people, more than a billion smartphones and over 800 million internet users, Chinese companies have unrivaled amounts of data with which to develop and test their algorithms. As high-profile venture capitalist Kai-Fu Lee so aptly puts it, if ‘data is the new oil,’ then ‘China is the Saudi Arabia of data.’"
Lam points to the raft of policies formulated by the Chinese government to push AI development and application.
"Since Beijing designated AI a national priority, billions of dollars of venture capital investment have been attracted," says Lam. "This has created hugely valuable startups such as SenseTime and ByteDance (owner of video-sharing social networking service TikTok) and sparked a talent war in the process."
Beijing is pouring vast sums into AI. The Chinese Ministry of Industry and Information Technology recently announcing plans to allocate around $950 million annually to fund strategic AI projects. Billions more have already been invested in startups and research programs. Municipal governments are getting in on the act too—the city of Tianjin has announced plans to set up a RMB 100 billion (around $14 billion) investment fund to stimulate development of the local AI industry.
Private investment in Chinese AI firms has also soared. According to Stanford University's 2019 AI Index Report, between July 2018 and July 2019, 486 Chinese startups received an average of just over $34 million per company, three times as much as their U.S. counterparts, and nearly eight times as much as European companies. While investment dipped in the second half of 2019, most experts attribute this to recalibration of the Chinese AI market, rather than the onset of a so-called "AI winter."
The talent dynamic
There are also signs that the advantage in AI talent held by the U.S. may be slipping away. In the past, many Chinese computer scientists studying at U.S. universities would simply stay on to work for tech multinationals.
"Between 1998 and 2017, the U.S. gained 1,283 foreign AI academic researchers, while China only gained 53," says Malmsten. "But this uneven dynamic is steadily shifting. China is already working on more initiatives to increase AI-related education at the university level, while increasingly competitive salaries are enticing Chinese IT professionals to return home from overseas."
A recent report by the Center for Security and Emerging Technology at Georgetown University found that the tightening of immigration policies in the U.S. is also proving counterproductive to maintaining the lead in AI talent.
"Nearly half of all Silicon Valley professionals are either Chinese or Indian," says Malmsten. "There have been complaints from Silicon Valley companies that visas are now the largest bottleneck on talent."
"Despite the U.S. and China being pitted as AI enemies, companies and institutions continue to reach cross-border for collaboration and mutual gain."
Challenges and opportunities
As it looks to build a $ 150 billion AI industry over the next decade, China will have to overcome a number of fundamental challenges.
"There are three areas where China needs to make significant progress—hardware, research and algorithms," says Lam. "Of these, the production of hardware such as microprocessors and chips remains the biggest barrier."
Despite decades of effort and huge investment in chip research, today China still imports 90 percent of its chips. U.S.-based company Nvidia is currently in a league of its own when it comes to the development and manufacture of graphic processing units, the most widely used chip for supporting AI algorithms.
Yet the fact that AI chips handle data differently to traditional silicon logic circuits has all but leveled the playing field in terms of chip development. Leveraging their unrivaled access to data, Chinese companies may even have the edge when it comes to designing the hardware required to host cutting-edge algorithms. Indeed, both Huawei and Alibaba introduced AI chips in 2019, with the former claiming its Ascend 910 chip "has more computing power than any other AI processor in the world."
While China has led the world in the number of first patent filings in AI over the last few years, the country's "patent first, innovate later" culture and top-down subsidy system is reflected in their generally low quality. By contrast, AI development in the U.S. is market-driven, meaning U.S. patents are far more frequently connected to commercially viable applications.
Yet, Malmsten says the quality of Chinese research is now improving, with dedicated research centers cropping up across the country, financed by the state as well as institutions as diverse as Tsinghua University and Google.
"China has certainly focused on quantity over quality in AI research to date," says Malmsten. "Yet a study conducted in 2019 shows that China's output of influential AI research papers is increasing and will soon overtake that of the U.S."
Cross-border commercialization
China's quest for AI supremacy is more than a simple race to the top against the U.S. Cutting-edge AI technologies can underpin advances in industries such as health care, transportation and communications, with pioneering countries steering developments and reaping the economic rewards.
While much is made of China's focus on developing AI for surveillance purposes, today Chinese AI developers are making rapid progress with the application of AI for a whole range of industrial and commercial purposes.
China's trio of tech titans—namely Baidu, Alibaba and Tencent (known as the BAT)—are now bringing their AI capabilities to bear at home and overseas, as they jockey for position in the new AI world order. Backed by the recruitment of U.S. talent, they are forming global partnerships to advance everything from smart city solutions and autonomous driving, to conversational AI and predictive health care.
"Despite the U.S. and China being pitted as AI enemies, companies and institutions continue to reach cross-border for collaboration and mutual gain," says Malmsten.
Over the last few years Baidu has snapped up a number of Western AI startups, such as xPerception and Kitt.ai, as it looks to boost its own AI capabilities.
"Although they are becoming increasingly discerning in their purchases, Chinese tech giants and investment firms continue to show strong interest in western AI companies," says Lam.
Health care focus
Health care AI is one of Tencent's main priorities, with the company's efforts underpinning China's huge push to be a world leader in genomics and personalized medicine. As China's biggest social network company, with 1 billion users on its WeChat app, it has unparalleled access to health care-related data for training AI algorithms (more than 38,000 Chinese medical institutions have a WeChat account and 2,000 Chinese hospitals accept WeChat payment).
Through investments and partnerships, Tencent is now bringing health care AI technology from around the world to China. In 2018, the company teamed up with London-based AI health care startup Babylon to deliver health care services to WeChat users. Established in 2013, the British firm has developed an AI-powered digital platform that helps users find information on a range of medical issues and conditions.
Babylon founder and CEO Ali Parsa sees myriad opportunities for Western AI-centric health care service providers in the Chinese market. "China's health care system is now facing a range of challenges, from an aging population and a burgeoning chronic disease burden, to a shortage of quality medical resources and ballooning public medical expenses," says Parsa. "AI has the potential to be a game changer by saving time and money, and improving the diagnosis, management and treatment of disease."
Future focus
Led by companies such as Huawei, SenseTime, Baidu and Cambricon, and driven onwards by a supportive national policy framework, buoyant investment climate and the needs and desires of a massive consumer market, China's AI industry is gradually shifting its focus from research to commercialization. Regardless of targets and international competition, the next few years will offer lucrative opportunities to those armed with the right technology and intelligence.