By Bonnie Buchanan1
1Head of Department of Finance and Accounting and Professor of Finance, Surrey Business School, University of Surrey,
Artificial intelligence is impacting industries ranging from IT to financial services, healthcare, education, surveillance and regulation. Approximately 5154 AI startups have been established globally during the past five years.1 A 2019 PwC report2 estimates that AI could add as much as $15.7 trillion to the global economy by 2030, and claims the greatest gains will be experienced in China and North America. Between 2012 and 2016 the US invested $18.2 billion into AI compared with $2.6 billion in China and $850 million in the UK.3
There are two main explanations which account for AI’s rapid growth. Firstly, exponential advances in computing power have led to declining processing and data storage costs. Secondly, data availability has increased on a massive scale. AI is now a national priority, but approaches vary. China, the United Kingdom (UK) and the European Union (EU) have adopted a government-led approach to AI. The United States’ (US) AI strategy has been dominated (and self-regulated) by big tech companies like Microsoft, Facebook and Amazon. Singapore’s AI structure emphasizes a more “human-centric” approach that includes explainability, transparency and fairness to establish public trust in AI. The Chinese strategy is also based on a very different financial market model. Whereas the economies of the US, UK and Singapore are based on an English common law model (which emphasizes strong shareholder and creditor protection), China’s market model is best described as being in the primary stage of socialism, which is to say, the state controls access to capital, influences investment decisions and stock market listings and thereby combines state power with capitalist tools. At the local government level, there are financial incentives to encourage AI-related innovations.
In 2017, the Chinese government announced the “Next Generation Artificial Intelligence Development Plan” with the goal of China being the AI world leader by 2030. The Chinese government presented a timeline where it expects companies and research facilities to be at parity with the US by 2020.4,5 In 2017, the Chinese government also announced plans to set up an “intelligence industry zone” near Tianjin to support the AI industry. Historically, the US has dominated the AI landscape, but this pattern is now shifting. There were 3033 AI startups in the US between 2000 and 2016, accounting for 37.41% of the global total. Since 2016 the proportion has decreased, dropping to under 30% for the first time.
In 2017, China surpassed the US in terms of AI startup funding,6 accounting for 48% of the global total. The US is also losing its global AI equity deal share, decreasing from 77% to 50% of equity deal share over five years.7 During the same period China accounted for 68.67% of Asian AI startups and corresponding AI funding was 60.22% of the Asian market. Many Chinese cities and provinces dominate other Asian countries. In terms of the number of AI companies, there are 1387 in Beijing, 792 in Guangdong and 154 in Shanghai, compared with 57 in Singapore and 283 in India.8 Beijing has attracted $1387m in AI funding, followed by Guangdong ($792m) and Shanghai ($154m), the total exceeding that of Japan ($436.81m) and the UK ($1251m).9 Between 2012 and 2016 the compound growth rate of AI patents was 33.2% per annum. Currently the US and China hold over 50% of all AI patents (35,508 in US and 34,345 in China, respectively) but patents are being filed at a much faster pace in China. The US holds 32% of machine learning patents and 26% of natural language processing patents, followed by China with 23% and 14%, respectively.10
Currently China dominates the machine vision patent category (55% of 150,000 patents globally). Machine vision describes object and facial recognition and is useful in public security, healthcare, e-commerce and autonomous driving. China outpaces the US in terms of both deep learning and AI-related patent publications (by a factor of more than five), and the gap is closing for machine learning-related patent publications. Cross-border investment is on the rise but not in an equal manner. There are now more Chinese investments in AI startups in the US, rather than vice versa. What explains the increasing dominance of China in the area of AI? Firstly, many machine learning techniques require vast amounts of data, and China has the scale. China’s online population of 730 million people is almost twice the size of the US, with the WeChat platform alone accounting for over a billion users. Secondly, there are two prominent technologies fuelling the drive, namely facial recognition and AI chips.11 Both the US and China compete heavily in AI chip technology. Chinese corporations such as Baidu, Tencent, iFlytek and JD.com invest heavily in AI, both domestically and abroad.
Chinese and American “big tech” firms also differ in terms of their AI focus. Microsoft, Google and IBM focus on machine learning, speech recognition and speech synthesis, whereas Tencent, Alibaba and Baidu focus on AI searching and facial recognition.12 Deep neural networks applications need to supplement central processing units (CPU). AI chips draw on graphic processing units (GPU) technology which is then applied to AI, machine learning and deep learning problems.
From a corporate perspective, Alibaba is aiming to have its first AI chips on the market in 2019. Ant Financial already uses facial recognition for payments at Alibaba-owned retail stores. In 2016, Ant Financial, Foxconn and the city of Hangzhou partnered for the “City Brain” project using AI data from social feeds and surveillance cameras. Additionally, 55 cities participate in the “Sharp Eyes” project whose surveillance data could end up in the nation’s Social Credit System, a measure to gauge citizens’ “trustworthiness”. Three big techs, Baidu, Alibaba and Tencent, are also privy to what consumers buy, where they travel and who they chat to online. In fact, Baidu has shifted its business strategy from “mobile first” to “AI first”.
Internationally, there is progress to adopt a more coherent cross-border AI governance network. In May 2019, 42 countries signed an accord which committed to common AI principles. Even though China did not endorse the principles, it is developing its own framework. The UN, OECD and Council of Europe have all formulated their own AI goals. In April 2019, the European Commission published AI guidelines. Growth in the EU AI market has been dominated by the Nordics, France and Germany, but overall is more fragmented because of heterogeneous resources and regulation.
Today, global political leaders have choices to make. It will be interesting to see which countries continue to adopt an inward-looking, and almost protectionist, nationalistic stance with regards to building AI capability compared to other countries which may be more inclined to leverage the existing institutional framework for international cooperation. Regardless of the stance adopted, the geopolitical landscape is sure to be heavily impacted by national AI strategies, although the future consequences of today’s decisions are far from clear.