At what point did China decide that it wanted to invest heavily in AI development, and what has been driving this process within China?
Jinghan Zeng: Since the 1980s the Chinese government has maintained a keen interest in e-government as a way of improving its governance capacity. Since then it has sought to use cutting-edge digital technology to better equip the civil service and has expended a lot of effort to learn from the technological innovations of the West. Over time China has adapted these innovations to its mode of governance, from the development of the ‘Great Firewall’ to incorporating the power of big data into bureaucratic processes. And now, more recently, we are seeing attention turn to the potential of artificial intelligence.
2017 is the pivotal year in China’s ambition to become the AI global superpower. Before this there were efforts to invest in AI at the provincial government level or in discrete industrial sectors, but they were minimal. In 2017 the Chinese government published its own ‘New Generation AI Development Plan’ which set out a three-step strategy to meet its aspirations for this technology. The first step was to achieve the same level of competency as leading countries, such as the United States, and to develop an industry worth at least RMB 150 billion by 2020. The second step is to become the country which regularly makes the major breakthroughs in AI, and develop an industry worth at least RMB 400 billion by 2025. For the final step, to be achieved by 2030, China aims to become the leading AI power with an industry worth at least RMB 1000 billion. The announcement gained a lot of traction and provided impetus to capital markets within China to invest in these technologies.
Now, of course, this turn to AI is not solely a Chinese one. Since 2015 the AI governance shift has become a global phenomenon with many countries developing their own strategic plans for seizing the opportunities AI can offer. In the United States, for instance, in early 2019 President Donald Trump signed an executive order to establish the ‘American AI Initiative’, a move that has been mirrored in almost 50 other countries. So an element of international competition on the AI question may also be driving China’s approach.
How successfully has this strategy been implemented?
Jinghan Zeng: The 2017 plan was primarily a way of mobilizing domestic actors to push forward innovation; since then many provincial governments have jumped onto the AI bandwagon. More than 20 cities, including Beijing, Shanghai and Chongqing, have developed their own AI policies which aim to encourage companies to invest in AI development through a range of measures including tax reductions and entrepreneurship schemes. While this process may result in inefficiencies and the duplication of efforts, there is little doubt that since 2017 AI development has become a national public concern.
In 2019 there was an overall fall in investment in AI and the COVID-19 pandemic has also put pressure on the Chinese economy, so it is possible that the ambitious targets set within the 2017 strategy may not be reached on time. However, as evidenced by the widespread adoption of AI technologies by provincial governments across China, the transition towards so-called ‘intelligent government’ has continued despite the economic slowdown.
One example of this can be found in Guangdong, where the local government has used facial recognition technology to drastically streamline bureaucratic processes. Where before the application process to receive a commercial business licence took at least three days, an AI-driven approach to identity verification has reduced the wait to just ten minutes. Similar innovations have also sped up processes in the banking and transport industries.
Elsewhere, police forces have demonstrated the value of AI in the fight against child trafficking, where technology is used to create detailed images of how a missing child might look five years later. This example has been prominently promoted by the Chinese government as a way of demonstrating the positive influence of AI. While the privacy trade-off is rarely acknowledged, the benefits in terms of convenience and security are constantly emphasized.
Why does China’s socio-political context make it uniquely well-placed to seize the opportunities afforded by AI?
Jinghan Zeng: The first thing to note is the Chinese political system. China has been leading in the development of, for example, AI-driven facial recognition technologies partly because there are fewer legal barriers to this sort of work than in the United States or United Kingdom. Of course, on some level, the security services in those countries have been experimenting with facial recognition, but this has been heavily criticized by society at large. The COVID-19 crisis has provided crucial momentum for China’s AI strategy. As has been demonstrated across the world, in order to prevent the transmission of the virus you must have tight controls on the movement of the population. Countries like China and South Korea have pioneered different kinds of smartphone apps to identify the movements of their populations. In the West, this form of contact tracing was accompanied by public debate about the use, processing and storage of that data. While in the UK, for example, we have seen much public criticism of the government on this question, in China people have received contact tracing technology far more positively, seeing it as a necessary trade-off in exchange for improved health outcomes.
Public debates have not shaped the national agenda on AI within China. The government has a strong will to strengthen its control over society by using these technologies and, relative to the West, there is less social resistance to practices such as mass surveillance. There is some concern voiced over privacy or abuse of data, but these concerns are directed at how the private sector uses AI, not the government. That’s a critical difference: the government has attempted to highlight positive AI stories in order to shape public perceptions of the technology. Additionally, Chinese public-private sector relations are also very different. Major corporations such as Alibaba or Baidu have to work very closely with the government in order to continue to conduct this sort of technological development, whereas US tech giants have more autonomy over how they conduct business, Apple’s continuous refusal to cooperate with the FBI on unlocking phones being one example. There are serious adverse consequences that incentivize Chinese public-private collaboration.
Can we expect to see China successfully exporting these new technologies? Or will privacy and security concerns prevent this?
Jinghan Zeng: In the past certain countries have been keen to adopt Chinese governance technologies, for example Russia and Iran’s interest in the ‘Great Firewall’ system. And I think that from a technological point of view, leaving aside the politics for now, Chinese-developed AI governance tools could well be attractive to buyers in other countries. However, if you bring in politics and ideology then that is a different matter. I cannot see how countries in the West could import Chinese systems or apps, not least because of national security concerns such as those we’ve seen with Huawei and 5G.
But as we have discussed, China is uniquely well-placed to develop advanced AI technologies, given the vast population they can draw on to provide data to pursue the machine learning model. So, while for political reasons we may not see these tools adopted by European or North American states, we could absolutely see governments which are more willing to sacrifice individual privacy for greater control over their populations take an interest in exchange.
What are the risks for the Chinese government in implementing this AI strategy?
Jinghan Zeng: The first thing I think we have to understand is why China wants to become an AI superpower in the first place; the answer to this is primarily economic. The Chinese government derives much of its domestic legitimacy from boosting economic growth and raising its population’s living standards. The typical approach to achieving this is through mobilizing domestic actors at the national level with high-profile announcements and ambitious targets, such as those included in the AI strategy. In contrast to the western market-based approach to economic innovation, however, this Chinese model risks considerable wastage and inefficiency, and puts the burden of innovation on the government’s own resources. If the strategy fails to achieve its aims then the economic costs could be considerable.
Beyond this efficiency concern, there is the much larger question of the social transformation which AI could instigate within China. For instance, very little thought has yet been put into the consequences of mass unemployment due to AI-enabled automation. A recent McKinsey report predicted that by 2030 between 400-800 million people across the world will be forced to find new professions due to automation. As China’s population accounts for approximately 12% of the global workforce, they could carry the heaviest burden of this transformation. Above all, the Chinese Communist Party values social stability, but the potential disruption to the existing social order that AI promises has not yet been addressed. Some argue that AI will create a world where labour is no longer the responsibility of humans, but even if that vision is realized further questions remain about how to pursue a meaningful life without work. By aiming for AI superpower status China is accepting the risk that their society will bear the brunt of any downsides to AI experimentation, while potential competitors in the West take a more circumspect route towards the adoption of these technologies.
Looking ahead, do you think China will realize its ambition of becoming an AI superpower?
Jinghan Zeng: Only the United States and China have the resources to compete for predominance. China’s advantage can be found in its vast internal market, which provides developers with unparalleled datasets with which to train AI technologies. But the United States has a considerable head start in developing the technologies themselves.
The Chinese state approach towards AI may contribute to the fact that China lacks genuine innovations in the field, especially those that are appreciated by the market. Questions remain about the extent to which China’s AI investment can turn into something that the global market will buy into. This is one of the key reasons why investment in China’s AI industry declined in 2019.
In addition, China’s open AI ambition has alarmed its near competitors, especially in the US and Europe. Here, geopolitics including the current US-China tension and China’s nationalistic tone is unhelpful to China’s AI progress. Global cooperation in AI is vital to ensuring that these technologies are used in ways that benefit humanity. The geopolitical competition driving AI development is an unwelcome impediment to realising this, and could hamper China’s superpower intentions.
This interview draws on Jinghan Zeng’s article, ‘Artificial intelligence and China’s authoritarian governance’, which was published in the November 2020 issue of International Affairs.