On a recent trip to Hong Kong and Shenzhen, which included chairing the AI Summit in Hong Kong and meeting with several companies, I got to see and hear firsthand how China is emerging as one of the powerhouses for the development and adoption of artificial intelligence (AI). While in Europe and the United States the AI news is largely focused on how Silicon Valley, Montreal, London, Paris, and Berlin are becoming hubs for AI, the visibility into Chinese AI activity is fairly limited. Recently, there has been increasing focus on China in the press and its ascent, but nothing beats going to the region and getting a front row view.
The Factors behind China’s Success
Several factors are helping China and its rise as an AI powerhouse.
- Academics: Chinese universities, including those in Hong Kong, have a very strong focus on math and science and are now starting to play an increasing role in the AI conference circuit, winning the ImageNet competition for the last few years. Tsinghua University, The Chinese University of Hong Kong, Peking University, and Nanjing University are some of the top ranked universities that regularly show up on the list of accepted papers across the International Conference on Learning Representations (ICLR), the Conference on Neural Information Processing Systems (NIPS), and other key AI conferences.
- Government: The Chinese Government is promoting AI development. A recent report from the State Council stated that AI is expected to generate $147.8 billion in output by 2030, with the goal of becoming a global leader in AI. It is one of the most wide-ranging initiatives from any government seen so far and spans funding, integration of AI into traditional industries like agriculture and manufacturing, application of AI into public services, laws and ethics formation, and many other aspects. The State Council has called for Chinese businesses, universities, and defense institutions to work closely in developing AI, and to integrate civil and defense initiatives. China’s AI initiatives will span robotics, augmented reality (AR), and virtual reality (VR). There is also a suggestion to create “AI innovation pilot zones” where companies could experiment with new AI applications, representing a sandbox for AI use cases.
- Startups: The venture capital (VC) and startup ecosystem in China is heavily invested in AI. The Tencent Research Institute estimates that China-based AI firms achieved cumulative total investment of CNY63.5 billion (US$9.5 billion) as of the end of June 2017. This is still lower than U.S.-based AI investment, which was estimated to be US$14.2 billion, but is significantly higher than any other country. China is estimated to have 23% of the AI companies, with the United States accounting for 42% and the rest spread out across Europe, Japan, India, Israel, and others. iCarbonX and SenseTime are two high-profile companies that have raised significant rounds of investment recently, with iCarbonX raising more than $200 million and SenseTime raising $400 million, both setting records for any AI startup.
- Data: Chinese internet giants Baidu, Tencent, and Alibaba are at the forefront of AI development as they service one of the largest internet user bases in the world, estimated to be more than 730 million at the end of 2016, with more than 95% of that user base using smartphones. Large user bases that share data with their service providers are the best fuel for AI algorithms, and while Google and Facebook can boast 1 billion users on their platforms, those users are spread out across the world in multiple jurisdictions with varied privacy laws. The recent European General Data Protection Regulation (GDPR), which will come into force in 2018, could make a massive dent in that user base for U.S. internet giants, especially in relation to using the data for training AI algorithms. Chinese AI companies have no such privacy issues to worry about. For example, SenseTime, the facial recognition startup that has raised more than $400 million, used government data from cities like Guangzhou (with a population of 14 million) and its closed-circuit television (CCTV) data to train its algorithms. While DeepMind in the United Kingdom has run into trouble accessing health data from the National Health Service, companies like iCarbonX have ambitions to hold the genomic data of 1 billion Chinese to provide them with personalized health services.
- Scale and Speed: Anyone who has been to China understands that the Chinese are very efficient at deploying and implementing things. Whether it is infrastructure or government policy, there is a diligent focus on getting things done. Startups like SenseTime, which develops face recognition technology, and has only been around for a couple of years has already gotten multiple cities, including Shanghai and Beijing, to deploy its technology for surveillance systems. As for technologies like facial recognition, China is leading the world in terms of deployments and, apart from SenseTime, startups like Megvii, Alibaba, and Baidu are all competing in this area. Megvii’s Face++ technology is now a default face recognition platform for internet banks and financial services across China. In terms of voice recognition technology, companies like iFlytek and Baidu have raced ahead in making major advances in reducing the error rate for recognizing Mandarin, with Apple’s Siri falling behind. iFlytek is now being used in government courtrooms for transcribing court cases, and by thousands of companies using the technology in robotic toys, intelligent buildings, call centers, mobile voice assistants, and education applications. In healthcare, a recent competition announced by the Fuzhou City government made 80 exabytes of heart ultrasound videos available to companies to build an AI tool that could identify heart diseases more accurately than human doctors. The deadline offered for the competition was 3 months.
On Course to Win the Artificial Intelligence Race
China clearly has many advantages when it comes to AI: a growing presence of Chinese institutions in AI research, access to vast datasets, almost limitless financial muscle, a tireless work ethic and system that values efficiency, and a government that understands the importance of AI and its role in shaping future societies. While the United States is leading with innovation and academic breakthroughs, its current political establishment does not place AI as a top priority. Similarly, in Europe, there is a very strong focus on AI research and a growing ecosystem of startups; however, we have not seen any major commitment from governments, as seen in China. We are entering an arms race for AI, and China is on course to win it, unless Western governments or countries like India start taking measures to consider AI as a top priority.