Chapter 3. Country case studies
3.3 China
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
infrastructures to cope with low birthrates, aging society, and the socioeconomic impact they entail (Mashiko, 2020).
Alongside legal changes and liberalization in data-processing and utilization, a law came into effect in 2018 to designate geographically defined National Strategic Special Zones in locations across the country with tax breaks for resettlement and relaxed or discarded regulations that allowed companies to carry out government-approved pilot projects in regulatory sandboxes, that is deregulated test environments to try out new business models and generate necessary data for further research, commercialization, or nationwide usage, including permissions for robot-human interactions in the streets in Fukuoka, or AI-automated robot taxis in Fujisawa City for disaster or life support services (Shimpo, 2018). Upon success, deregulation could be expanded across the country. The sandboxes do not only provide the space for innovative thinking and creative failure, but they also strengthen structurally weaker regions outside the big metropolitan areas.
3.3 China
3.3.1 Conditions for data-derived value
In 2015, China piggybacked on the idea of the 4.0-concept and hence introduced its Made in China 2025 industrial upgrading strategy. However, the country’s industry is more heterogeneous than Western economies’ in terms of its technological capacities.
There are billion-dollar corporations with high technology and many small and medium-sized enterprises that have not yet reached levels of the third industrial revolution (Ferenzy, 2018). Therefore, China’s 2025 strategy aims to achieve a large-scale improvement of general automation and notably its competitiveness in the production and manufacturing – a hint towards maintaining certain levels of the
“factory Asia” model but with ambitions to catch up in core technologies and move towards indigenous “value-added factory Asia” (Kam, 2017).
Against the demographic backdrop of the world’s largest population and growing internet penetration rates, the potential for a vivid digital ecosystem is undoubtedly high and complemented by a state system that allows the government targeted data collection and usage for systematic analysis and information extraction to an extent greater than anywhere else in the world. Despite ranking very low for availability and quality of public data (Lee, 2018), its sheer size in population and
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
landmass provides China with an advantageous position in generating data,
developing as well as applying data-driven solutions to a variety of settings. With 840 million internet users in 2019 and predicted increase to 975 million by 2023 (Statista, 2019), all of which already generate a massive data volume that concentrates in a few major Chinese technology companies, namely the ecosystems around Tencent and Alibaba. China’s rapid socioeconomic transformation has helped many Chinese companies like these to thrive in environments of a growing Chinese middle-class with increasing income and the purchasing power to satisfy consumption aspirations and pent-up demands for financial services (Lee, 2018; Naughton, 2018). Moreover, the transformation occurred all amidst the proliferation of internet technologies and mobile phones, which turned out to be the major factor behind the success of many Chinese big tech companies who could reach out to millions of people at a very low cost and channel their resources into generating innovative forms of namely financial technologies (fintech) and e-commerce (Ferenzy, 2018). Moreover, the rise of
Chinese tech companies is largely due to capable technocratic policymaking and governmental regulations (Naughton, 2018). As for fintech, for instance, regulations were relaxed to an extent that these companies could leverage the potential of internet finance without greater impediments in tandem with protectionist policies as a shield against foreign competition, which gave them scope to grow into “National
Champions” who enjoy subsidies and government-backed investments (Ferenzy, 2018).
In spite of its semiconductor industry base lagging behind in terms of precision and state-of-the-art chip production, “China is adopting a ‘catch-up’
approach in the hardware necessary to train and execute AI algorithms” (Ding, 2018, p. 4). It now hosts 208 out of the world’s most powerful 500 supercomputers, the largest amount of supercomputers in the world, with Sunway TaihuLight and Tianhe-2A in Wuxi and Guangzhou, respectively, among the top ten (Strohmaier et al., 2019).
Also, as of 2018, China ranked with 1,011 out of 4,925 monitored AI-enterprises second after the U.S. with 2,028 (Tsinghua University, 2018, p. 46).
3.3.2 Institutional framework
Since 2013, China has released various policies addressing issues around big data technologies and AI, including the 2013 State Council Guidelines on Promoting the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Healthy and Orderly Development of the Internet of Things, followed in 2014 by the State Council Notice on Issuing Made in China 2025, the State Council Guidelines on Promoting the “Internet+” Action, and the State Council Notice on Issuing the Action Outline for Promoting the Development of Big Data. Thereupon in 2016, the
Thirteenth Five-year Plan on National Economic and Social Development, and State Council Notice on Issuing the “Next Generation Artificial Intelligence Development Plan” were released in 2017 with specific target development directions and priority areas of China’s AI development (Tsinghua University, 2018).
For instance, immediate guidelines for the industry stating to put AI at the core of socio-economic development throughout the digital transition were released in form of the 2016 Chinese Three-Year Guidance for Internet Plus Artificial Intelligence Plan and backed by the Three-Year Action Plan for Promoting Development of a New Generation Artificial Intelligence Industry. The latter specifically sets out the need for an advanced public support system to drive growth, innovation, and breakthroughs in AI, but in return prompts companies to accelerate IoT/IIoT development including network hardware and software to create a next-generation internet infrastructure for smart factories and smart devices capable of video image identification or service tasks (Ding, 2018; Sirui, 2019). The Ministry of Industry and Information
Technology (MIIT) is the respective body in charge of overseeing progress and make adequate adjustments.
The new national AI development strategies represent an integral part of the national development strategy to realize president Xi Jinping’s Chinese Dream. AI-related strategies, such as Internet+ and the AI Three-Year Implementation Plan, are funded by the State Development and Reform Commission, the Ministry of Science and Technology (MOST), the Ministry of Industry and Information Technology (MIIT), and the Cyberspace Administration of China. Moreover, the State Council’s AI plan provides for establishing a new office under the direction of the MOST to direct full responsibility for implementation. The State Council issued the guidelines and the MOST thereupon announced to launch the implementation of the National AI Development Plan with its objectives that involve the major Chinese tech giants as the chosen national heroes to assist in datafication and platformization: Baidu for an autopilot AI open innovation platform; Alibaba for building a smart city AI open innovation platform; Tencent for medical images AI open innovation platform;
iFlytek for intelligent voice AI open innovation platform.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Traditionally, partnerships involving strategic assets and sensitive information have been reserved for China’s state-owned enterprises, but policymakers have realized that rapidly advancing big data and AI capabilities are developed outside the scope of traditional state-owned enterprises.8 This approach to advancing digital infrastructures aims to advance leading companies and platforms to emerging as regulatory stakeholders in the process (Chen & Qiu, 2019). In doing so, the capabilities of private firms are tied closer to the Chinese government’s long-term visions and sphere of influence as articulated in the New Generation AI Development Plan. Digital governance is cast to big tech giants that get special roles but, at the same time, are expected to serve as platforms for others (Chen & Qiu, 2019; Larsen, 2019). The “Guiding Opinions of the General Office of the State Council on
Promoting the Healthy Development of the Platform Economy” (The State Council, 2019) were issued in August of 2019, emphasizing that the platform economy is a new way of organizing productivity and a new driving force for economic
development:
“It plays an important role in optimizing the allocation of resources, promoting cross-border harmonious development and mass entrepreneurial innovation, promoting industrial upgrading, expanding the consumer market, and
especially increasing employment. It is necessary to adhere to the guidance of Xi Jinping's thoughts on socialism with Chinese characteristics in the new era … while continuing to stimulate market vitality. Increased policy guidance … and the establishment of innovative monitoring concepts and methods are required … to adapt to the characteristics of the development of the platform economy and issues that may emerge as well as creating a fair market environment for competition.” (The State Council, 2019, paragraph 1, author's own translation)
The platform economy and AI as the underpinning technology for China’s
development objective to firmly establish China as a global innovation center in AI by 2030, emphasizing that all related industries should generate an output value of one trillion Renminbi, approximately US$150 billion (Ding, 2018; The State Council,
8 Yin and Li (2019) note that state-ownership –as opposed to state affiliation– is generally rather rare among Chinese Internet companies with only a few having shares held by the government; larger and globally-oriented private Internet companies, however, show more distinguished forms of government affiliation, not least because of preferential treatments they enjoy on higher and more resourceful administrative levels. Through this channel, the PRC government can still exercise their influence on Internet companies’ development and steer technological progress it deems necessary to excel digitally and globally.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2017). Also, development and gross output benchmarks have to be achieved incrementally by 2020, 2025, and 2030.9 At the same time, it serves the
comprehensive focus of intertwining technology, economy, society, and the army, thus, to improve productivity, social efficiency, and national security (Shi-Kupfer, 2019).
3.3.3 Regulation and commercialization
The sheer market size of China provides big players such as Alibaba, Baidu, and Tencent with great potential for AI applications on a large scale. This creates favorable conditions for a data-driven development of their business ecosystems in several ways. Firstly, the large-scale market is attractive for investors which shows through the fact that China follows second after the U.S. in terms of AI startups (383 as of 2017) with the worldwide highest funding in AI startups, in general, coming from within China itself before the US (Lee, 2018; Varadharajan, 2017). With regard to this, China’s current position as one of the leading global AI hubs is mainly due to private investments in application-oriented R&D. This funding pattern arose in tandem with the government’s efforts in bringing together companies and universities by cutting public funding to stimulate universities to go seek private third-party funding of research in the course of opening reforms that began in the 1980s (Chen, Sanders, & Wang, 2008). Thus, application-oriented R&D for commercialization was prioritized over basic research which the government in retrospect considered as a necessary means to catch up lay the foundations for future AI development and research. However, what is unneglectable is the fact that Chinese patents regarding AI and deep learning skyrocketed and overtook the US by far (Lee, 2018).
Application-oriented artificial intelligence systems are very likely to be
successfully commercialized in China, however, it is not so clear whether this success holds for international commercialization. On the one hand, Chinese companies enjoy
9 1) By 2020: core AI industry gross output > RMB 150 billion (USD 22.5 billion) AI-related industry gross output > RMB 1 trillion (USD 150.8 billion)
2) By 2025: AI industry gross output > RMB 400 billion (USD 60.3 billion) AI-related industry gross output > RMB 5 trillion (USD 754.0 billion)
3) By 2030: core AI industry gross output > RMB 1 trillion (USD 150.8 billion) AI-related gross output > RMB 10 trillion (USD 1.5 trillion)
(Ding, 2018, p. 10)
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
great benefits and advantages from a relatively highly protected domestic market with regard to global competition, and the Chinese government is furthermore planning to tighten its grip even more and increase control and influence on domestic big players in tech sectors, for example, through party committees to be formed in companies (Ding, 2018). On the other hand, a disadvantage might be posed through the quality of the data gathered and analyzed by Chinese businesses since they are very specific with regard to their locale and linguistic focus, which is enhanced by the censored intranet characteristics of the Chinese internet (Webster, Creemers, & Triolo, 2017).
In tandem with growing consciousness in the West regarding data privacy and data usage as pronounced through the European Union’s GDPR, it remains unclear whether Chinese methods regarding data-handling and usage can be applied in these markets to gather data and internationally commercialize home-grown solutions developed under Chinese data protection regulations.
To tackle questions arising from this ambivalence, the China Electronic Standardization Institute put forward ideas to create technology standards regarding AI applications and security-related issues in order to be at the forefront in
establishing a structure of global governance regarding AI development and application (Lee, 2018; Varadharajan, 2017). Therefore, the Standardization Administration of the People’s Republic of China is already a member body of the International Organization for Standardization’s (ISO) subcommittee ISO/IEC JTC 1/SC 42 which is “responsible for standardization in the area of artificial
intelligence … provid[ing] guidance to JTC 1, IEC, and ISO committees developing Artificial Intelligence applications” (ISO, 2019). Working Group 5, for instance, is solely led by a Chinese team and focuses on “standardization in the area of
computational approaches and computational characteristics of AI systems” (ISO, 2019).
To move beyond innovation and induce a true data-revolution for successful commercialization, China’s strategy seeks to not only promote AI applications but fundamentally change the economic ecosystem around them by modernizing its industrial base akin to industrial 4.0 revitalization endeavors in Japan, supporting new ones, and transform the way people integrate with them, for instance, open-sourcing state data should provide the data necessary to build platforms for further
enhancement of existing (cyber-)physical infrastructures and digital integration. For example, the AI platform City Brain by Alibaba Cloud aims to realize truly smart
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
cities, a daunting task but not unachievable if efficiently fed with public
state-provided data and complemented with further data sourced from, for example, Apollo by Baidu for autonomous driving, or Tencent’s AI platform for a smart public health system coupled with aforementioned Ping An Financial’s Ping An Good Doctor healthcare network (Ping An Healthcare And Technology Company, 2019;
Varadharajan, 2017). The latter has also reached out, providing Ping An medical services in overseas markets, namely the U.S., Singapore, and South Korea. Thus, the country has established itself with domestic brands and tech giants beyond
manufacturing that continue to stretch out and monetize on AI applications internationally, such as aforementioned Ping An, but also financial services. For instance, in East Asia, AntFinancial invested US$1.2 billion in South Korea’s most important mobile payment app called KakaoPay and in South East Asia, the first data-related buy-in by a Chinese firm was a US$1 billion investment by Alibaba in
Southeast Asia’s e-commerce and shopping platform Lazada that operates in six countries across the region (Russell, 2017a). Additionally, China’s Tencent and Alibaba strategically invested in and acquired complementary e-payment ventures in Thailand, Indonesia, Singapore, and the Philippines, amongst others (Ferenzy, 2018;
Lee, 2018). By investing, acquiring and, thus, placing itself at the top of Southeast Asia’s nascent startup ecosystem, China will most certainly play an influential and decisive role with the potential to develop and forge intra-Asian standards in data-based businesses. An argument can be made that this may particularly true for Southeast Asia as ASEAN-led and China-inclusive RCEP reflects Southeast Asian aspiration for economic integration with China, especially with regard to trade and data policies fit to emerging market conditions in ASEAN as opposed to imposed provision by industrialized economies of Asia or the West (Froese, 2018).
Although ethical norms are discussed and considered, a regulatory framework is to embedded into the law only between 2020 and 2030 (Ding, 2018). This enables political and corporate stakeholders to experiment with unfettered ways to extract and analyze data from the vast population and come up with unprecedented forms of applications. For instance, the social credit system that is planned to be implemented from 2020 is more of a tool for the communist party to exercise social governance and prevent socio-political conflicts rather than an economic innovation of data-fueled applications for economic purposes as seen by most liberal economies (Ferenzy,
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2018).10 On the one hand, it certainly is an innovative application that may bring social benefits to the broader population if the Chinese socialist perspective deems people’s upheaval and free opinion as interfering with the greater good of the country.
On the other hand, ethical issues are increasingly gaining attention as ‘morality’ is a rather vague and certainly no ideological term that its definitional variety should not be forced upon the population through a public ranking system with benefits and punishments depending on ideological benchmarks. Despite a culturally distinguished definition of Chinese from Western morals, the Chinese government acknowledges common-sense principles to be part of the developing process of algorithms and data-applications that are supposed to translate into people’s everyday lives for the sake of the social fabric, people’s welfare, and their economic viability (Knight, 2019; Shafto, 2016; Shin, 2019).
Ding (2018) demystifies views on China through the Eurocentric or Western lens through which China is often depicted as the loner that does not want to integrate in a global (Western) world order in that there was little to no discussion of issues of AI ethics and data safety in China due to an exploitative government neglecting its citizen’s rights. However, there is a debate on safety and ethical norms regarding AI, however, Ding (2018) adds that the Chinese approach to AI regulations and economic exploitation would certainly display ideological features that distinguish from
juridical frameworks set out by the U.S. or by Europe’s GDPR reflecting historically-derived different values (Webster et al., 2017). In 2019, the World Economic
Forum articulated its own AI principles in collaboration with scholars, business leaders, and policymakers from the U.S., China, and others, with Lee (2018) as prominent AI investor and researcher who was involved in establishing Microsoft’s and Google’s outposts in China. He deems Chinese AI norms very similar to Western ones. For instance, the MOST-affiliated Beijing Academy of Artificial Intelligence articulated the Beijing AI Principles in the beginning of 2019, setting out norms for scholarship and AI development, including the need for “human privacy, dignity,
10 The social credit system is designed to rate and rank the financial and social behavior of each citizen and legal person (which includes every company or other entity) in China, impacting access to not just credit but a broad ecosystem of punishments and rewards, including travel permission, domestic and international plane tickets, blacklists for employment at state-owned enterprises, rankings on dating apps, and discounts on utility bills (Ferenzy, 2018; Mahrenbach, Mayer, & Pfeffer, 2018; Yee, 2017).
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
freedom, autonomy, and rights” (Knight, 2019). This is in line with the assumption that China is interested in international commercialization facilitated by complying with the global common sense based ethical dimension of data-derived applications.