After a long period of “winter” for artificial intelligence, in 2017 it suddenly began catching increasing attention. Canada’s pioneering in adoption of an AI strategy, and a follow-up in the same year by Singapore, Japan, and China kept the “AI strategizing” ball rolling—culminating with the US adopting its “American AI Initiative” in 2019. As noted by Foreign Affairs, today, “the idea of an artificial intelligence arms race between China and the United States is ubiquitous”. However, there’s an increasing danger of a digital split between East and West. This is due to China’s historic efforts of advancing its sovereign internet, recent technological decoupling with US with potential “tit-for-tat negative reciprocity in a “race to the bottom” and a potential split between two diverging technological universes, when countries emulate the regulations of either the US or China.
The roots of the digital split lie in China’s continuous policies of “internet sovereignty”, with China essentially creating a system of “two internets”, paving the way for the digital split. Artificial intelligence might become the crucial leverage to deepen the digital split as it is already transforming economic structures and is likely to cause societal changes—even potentially causing the “ end of capitalism”, as argued by Feng Xiang, legal scholar at Tsinghua University. “Proximate” countries might adopt the Chinese approach, aggravating the digital split. The first signals are already here. In Vietnam, internet control seems to follow China’s footsteps. In Africa, China has been consistently transforming Africa’s information space. For partner countries of the BRI, Chinese companies have been providing facial recognition systems.
US and China: Divergence in AI?
Recently, the US has been an outspoken leader in the development of the emerging technologies, granting it power to almost unilaterally set up related market and regulatory standards and rules. However, the US has been increasingly challenged by China both in development of these emerging technologies and standard-setting—consider, for instance, iFlytek, one of the leaders in intelligent voice technologies, and China’s intensive government activity in internationally promoting AI standards developed by the China Electronics Standardization Institute.
Despite their policy of internet sovereignty, China’s innovation environment for decades has been converging towards the Western one following the catch-up logic of economic policy. Since 2003, however, China’s industrial policy has transformed into a “techno-industrial” one with a larger control of the centralized forces and emphasis on Indigeneity and self-sufficiency albeit with more participation of market forces, primarily technological giants such as BAT (Baidu, Alibaba, Tencent) and Huawei. US industrial policy with Trump’s administration taking over has also transformed. It is still predominantly horizontal with the government supporting the business environment and heavily investing in R&D. However, its trade component stopped being “free”, instead becoming “strategic” with the government through tariffs limiting other foreign companies’ access to American tech or through sanctions limiting American companies’ motivations to share technologies or invest in unfriendly countries.
Do changes in industrial policy in the US and China result in more support for AI?
In the US the primary beneficiaries of AI-related policy according to Trump’s AI Executive Order are national firms and workers. In the regulatory domain this Executive Order sets a market-driven approach, removing regulatory barriers “to enable the creation of new AI-related industries and the adoption of AI by today’s industries”. To date, mostly non-binding regulations of the executive branch are providing industry-specific guidances, informing firms on government stance on specific topics and even welcoming firms to work on removing regulatory barriers. However, legislatures are also strong players trying to promote the interests of citizens by tackling the potential negative implications of AI with Congress intensively introducing legislative initiatives, promoting issues of accountability, transparency, privacy, nondiscrimination and job displacement.
This “executive-legislative” balance is further translated into the security dimension. The executive office’s traditional concern about national security with increased importance of the cybersecurity component is clearly represented in the AI strategy—the Executive Order states national security as the 2nd goal after economic growth and before improvement of life quality and AI.gov explains cybersecurity challenges of “robust and safe AI”. Legislatures on the opposite are promoting “private security”. California as a state and San Francisco as a city are leading the way by limiting or even banning the use of AI-related facial and voice recognition technologies in their usage by government entities for surveillance purposes.
In the economic realm the legacy of maintaining healthy internal competition prevails with the above discussed market-driven approach. However, AI seems to be the continuation of the general policy of the Trump’s administration shifting of the status quo, especially in relations with China, which echoes the trade-war rhetoric towards China.
Chinese policy-makers in its AI policy and corresponding regulations despite the widely publicized advocacy of individuals and the public good, are de facto supporting national champions. Consequently, China consciously leaves the AI industry almost unregulated with the prevalence of recommendations and planning documents over legislative initiatives.
Traditionally, special attention is paid to security issues. China’s “New Generation Artificial Intelligence Development Plan” (AIDP) strategy document states that China will “promote all kinds of AI technology to become quickly embedded in the field of national defense innovation”. Additionally, AI is becoming a key tool for the implementation of so-called “Digital Leninism”, when new capabilities in facial recognition and text mining of vast datasets of micro-level individual behavior will better enable monitoring, surveillance and punishment.
Economically, China is rapidly moving from a model of national champions to a model of public-private partnership (“National AI Team”). However, while this kind of endorsement can accelerate the AI projects of team members, it could impact market vitality, posing a challenge to industry latecomers. Geopolitically, China’s general economic interests and influence have grown across Latin America, Africa, the Middle East and Europe and support for AI seems to be a part of these related policies.
Institutionally, in the Chinese model of support for AI (see Table 1) market forces seem to have a slight upper hand, while centralization forces are somewhat maintaining their dominance but carefully experimenting with decentralization.
Surprisingly, countries are not divided by an abyss in their AI regulatory approaches. The only two observable differences are (1) balancing role of US legislatures in ensuring citizens’ interest and rights ; (2) China’s pronounced reliance on national champions. Institutionally, China even seems to be converging to the United States. Seemingly, there’s little space for innovation in the model of support and given competition pressure both countries are willing to rely on proven practices with a very careful experimentation. However, governance models might be similar but if they are not complementary and are not a part of the multilateral effort or bilateral cooperation, the market specificities will result in advancing several different technological systems at the same time.
Deepening the split. Russia the first to follow China’s AI model?
With a closer look at the “AI geography” (see figure 1), three main groups of countries can be distinguished by the scope of the AI regulation policy. China, US and EU are among key wide-regulation actors, Russia is also seeking to join this group. Despite all the differences and existing prerequisites for a split, it’s still unclear whether there’s evidence for deepening the digital divide yet. Except for global leaders, for now there are only 5 “narrow” players (Great Britain, Singapore, Japan, Canada and South Korea) advancing a certain aspect of the AI regulation. Other countries only demonstrate intentions to join the niche echelon.
Notes: Legend: Wide scale players – key players in the global AI market, implementing regulatory AI measures and participating in global standards development. Niche players- introducing or developing industry regulations. Late-comers- countries interested, or developing strategies, AI implementation policies. Interested players- showing active interest in AI at the state level.
Russia, on the opposite is showing signals of joining China’s “AI camp” with similar features such as national champions as beneficiaries, national security with “surveillance taste” and geoeconomic focus.
Russian Strategy for AI as a high-level document states the priority of AI benefiting citizens but the incoming Federal National Project for Artificial Intelligence will be a part of the “Digital Economy Program”, which clearly targets national champions.
In the security domain Russian AI Strategy clearly puts emphasis on national security and throughout the document pays considerable attention to cybersecurity issues. In the domain of “private security” the Strategy talks about citizens’ rights but doesn’t explicitly discuss privacy-related questions. Evidence gathered by some think-tanks on surveillance and so-called “digital authoritarianism” leads to a tentative conclusion of prevalence of national security over private security in Russian AI model.
Considerations on national security related to AI of Russian authorities have been initially defined in geoeconomic terms (“he who can establish a monopoly in artificial intelligence—we are aware of the consequences—will rule the world”, Vladimir Putin, 2017), and the tone of the Russian AI Strategy clearly echoes this trend. The internal competition aspect is discussed in the “support for competition” principle but is in fact neglected by the implementation model with national champions at its core.
As discussed earlier, for the technological convergence not only similar policy approaches but cooperation instead of competition should take place. Russian industrial policy has been increasingly keeping an eye on Chinese practices and recent tensions with US of both China and Russia led to an unprecedented growth of technological Sino-Russian cooperation.
The Chinese policies of the government-led capitalism started to be an influential benchmark for Russia first in policies related to special zones in late 2000s and recently in its import substitution and export promotion policy-making themes. Recent sanction pressure from the West, further led Russian authorities to adopt China-style policies of technological self-sufficiency. Further down the road in support for AI Chinese “open innovation platform” of the AI National Team seems to be attractive for Russian policy-makers, as Russian state-owned bank “Sberbank” will be the integrative actor in developing and implementing AI activities.
Sino-Russian cooperation in emerging technologies has recently soared. As discussed in a recent report from the International Cyber Policy Centre, on the state level, both countries have already determined the 2019–2024 China–Russia Innovation Cooperation Work Plan; President Vladimir Putin signed a decree, designating 2020 as the year of Russian-Chinese Tech Cooperation. On the private level, Huawei has already launched a research program in Russia, committed itself to building 5G networks and expressed an intent to develop AI ecosystem, while Alibaba and Tencent are widening cooperation with Russian companies to develop e-commerce and cloud services respectively.
To sum up, for Russia in modern geopolitical realities, a potential movement towards the Chinese AI development model from economic perspectives is obvious. Among factors that might alter the current trajectory are the decrease in sanction pressure, the resumption of technological cooperation with western companies, and the access of US capital to Russian innovation.
Notes: Legend. (1) Security: narrow – considering only national security, wide – considering national security and personal citizens’ security. (2) Economic benefits: narrow – considering only internal economic benefits, wide – considering internal economic and geoeconomic benefits.
Should the US establishment get scared? Well, at least worried. Potential consequences of more countries joining China’s technological universe might be fatal to the US’ technological leadership. Imagine the whole BRICS movement towards the Chinese model—not impossible, Chinese is already technologically active in these countries.
Detailed policy prescription doesn’t exist yet and it’s not needed. Instead, US policy-makers are likely to benefit from revising the course of its new “strategic” trade policy (from hard to soft) so as not to fuel the move to self-sufficiency in other countries. Reconsidering the areas of cooperation with pivotal countries, especially in the domain of emerging technologies, will further help to prevent the deepening of the digital split.
Subscribe to us to receive our China Focus Newsletters!
Read more about US-China Relations:
Latest posts by The Editors (see all)
- A New Era of Chinese Technology and Innovation - December 14, 2020
- Chinese Analysis of COVID-19’s Impact on the Economy - November 23, 2020
- The Changing Political Views of Chinese Students in the U.S. - October 16, 2020