Artificial Intelligence with Chinese Characteristics - Transcript

Intro [00:00:07]

This is the Nordic Asia podcast.

Joanne Kuai [00:00:11]

Welcome to the Nordic Asia Podcast, a collaboration sharing expertise in Asia across the Nordic region. I'm Joanne Kuai, your host for today. I'm a PhD candidate at Karlstad University in Sweden and an affiliated PhD with the Nordic Institute of Asian Studies. Joining me today is Jinghan Zeng, professor of China and international studies at Lancaster University. His research focuses on China's domestic and international politics. We'll be talking about his latest book, Artificial Intelligence with Chinese Characteristics, National Strategy, Security and Authoritarian Governance. The book is the first book length study focusing on China's politics. Thank you so much for being here today.

Jinghan Zeng [00:00:56]

Thanks for having me.

Joanne Kuai [00:00:57]

So shall we begin by your definition of artificial intelligence? What is I? Would you please give us some examples? Well, I think this is a very good question. I think nowadays AI has almost become a slogan. To me, it is the buzzword across the world, including China. We should define AI clearly and seems to me AI is a very much umbrella term, referring to many meanings in different contexts. In my book, I meaning refer I to a wide range of digital technologies with ability to perform tasks that would usually require human intelligence. In that regard, there are three types of AI, which includes narrow generally and super. So narrow AI Also people would call it weak. AI refers to digital technologies with a narrow range of ability that is dedicated to specific tasks such as the virtual assistants in your iPhone, Siri or the room robots or self-driving car like Tesla. It is the most basic generation of AI and performs below human level. The current state of development actually all belongs to this narrow generation of AI. The general AI also called strongly AI is more advanced version that is yet to be achieved. It will have the ability that can perform as good as human intelligence and the most advanced type, which is super AI and which is still hypothetical, represents the most advanced generation of digital technology currently conceived. It will have the strong self-awareness and be able to surplus human intelligence in all area and in all the science frictions we talk about, many of them refers to actually super AI, and this definition is very much critical to my book and also to the Chinese context, because AI is not used as uniform concepts, it is used as a catchphrase. And to vigorously develop AI developed China into a superpower has now been used by the Chinese government as a policy slogan to mobilize domestic actors, and the concept proceed has leaving considerable room for interpretation. There is neither coherent understanding nor unified use of the concept of AI in China. Precisely because of this, a lot of Chinese concepts we need to understand very clearly AI, for example, an AI economy and the coal industry and all those kind of concepts appears in the Chinese document with specific numbers to measure them. It should be very careful regarding the definition of it. Otherwise then we are into some misunderstanding or confusion regarding those terms.

Joanne Kuai [00:03:42]

So the title of a book is Artificial Intelligence with Chinese Characteristics. Can you tell us what do you mean by AI with Chinese characteristics?

Jinghan Zeng [00:03:51]

Right. So the overall argument of my book, you see that China's AI approach is sophisticated and multifaceted. It's shaped by its domestic, political and economic environment, which has mixed implications to China. I call this Chinese approach AI with Chinese characteristics. Given my background on Chinese politics, I think socialism with Chinese characteristics is a good reference of it. And this refers to how Chinese concepts which adapt into China's conditions make it unique. And what is unique or what precisely are the Chinese characteristics? In this book, I mainly focus on three areas. The first, Chinese characteristics focus on the China's unique way of designing a strategy which make it different from United States. Many in the US has been actually characterized China's approach as a top down, geopolitical driven, national concerted approach. But my book actually saying this is wrong. My book really look into how there are strategic planning of AI from the top. At the same time, the how China's AI plan has been driven by contestation struggles for resources among domestic stakeholders who are actually economically motivated and have little awareness of. Geopolitical picture and how China's strategy is actually an upgrade of existing AI initiative to the national environment. I think China's strategy should be better understood as a broad policy manifesto. Accommodating competing interests of various domestic stakeholders to develop AI in China is a political slogan for mobilization, and this is different from how the strategy in the US and in Europe. And the second major difference and the characteristics is a way how China securitize AI. And usually when people are talking about securitization of how politician has been making some topic out of normal politics and make it a matter of security, they talk about this in a democracy, especially Western democratic society, but in China there's still a considerable needs to convince the domestic actors. So securitization of this AI to make AI national security matter in China's also a Taurean context is also very unique. In this book. I argue that China's government has been using securitization, trying to mobilize local actors, market actors, intellectuals and general public and this has been further escalated by the growing tension between China and the US. But while the securitization helped to achieve its goals, it also bring a very complicated implication to China as well. So that's the second major characteristics of this approach in China. And third one is obviously China's unique political system combined with its approach. China is a large loss authoring regime and the probably most politically resourceful regime who is now leading in AI and pioneering the governance of AI. So in this book, I talk about two things what happens when also in Value Match II, and b, what kind of ideology would a robot believe in? And also in context. And in this book I talk about different ways. How has been empowering also but same time bring considerable risk to the region as well.

Joanne Kuai [00:07:12]

Can you elaborate a little bit on the challenges and the potentials you just mentioned that I could bring to the authoritarian regime?

Jinghan Zeng [00:07:21]

So for the challenges regarding China's AI shortage, I think we have to look at the comparatively meaning in the US. Or if you read some US Congress report about China, they have been sensing some sort of jealous people. I'm not sure whether that might be the best word. People are jealous about China's whole of society approach, a nationally concerted strategy of that. And the meaning has been saying, you know, we should be doing the same in the US. We should integrate a whole of state approach to develop AI in the US in order to confront or in order to maintain American AI supremacy. My book disagree with this characterization of the summary of AI in China, but the book did talk about yes, the Chinese approach is a combination of state driven at the same time with the important role of local and subnational actors. And the benefit of that is very quickly mobilize domestic actors and draw attention on the metro. But the challenge of that is also the Chinese approach has the issue of low efficiency waste of resources. Because now, too, development has become a slogan. It has been manipulated by domestic, especially local and subnational actors, leading to considerable problems like efficiency of resources. At the same time, it has increased anxiety of China's competitors, leading to undesirable international environment. For China, the second major aspects of benefits and cost. Either way, China securities is a way Chinese government make it into a security matter. Obviously, by doing that, I think China's government successfully mobilized domestic actors to achieve its goals in this discourse to make people believe it is actually a security matter. And talk about the AI in a securities course. That's a way to achieve the government goal. But at the same time, it has also pushed by talking or firmly situated AI in the national security matter. In this book, I argue it has been pushed China's AI approach into an inward looking and nationalistic direction, which might bring a series of severe consequences for China's AI industry and leadership attention. Specifically in this book, I talk about how this make China less attractive to global labor and capital by producing the nationalistic environment, how it has been hybrid industry efficiency, by focusing on self reliance, or how it makes it harder for China to lead global governance, to further reinforce technological rivalry by negating the potential of global cooperation and how that constrains Chinese companies to global access. So overall, all of those could undermine China's key objective of. Boasting a booming economy and becoming a global leader in the third part of Australian governance and values. Obviously AI has been very useful to improve not only China's public services, but also to strengthen China's also authority in governance. It has been proving very helpful to the state surveillance, for example. However, I think the overall impact of AI on China's authoritarian rules and needs to look wider around surveillance and technological powers. I think it depends on three key pillars of political legitimacy. One is economic growth. Secondly, social stability. Third is ideology. In this book, I look into how I will work with authoritarian regime depends on whether AI is able to achieve or whether China is able to achieve a booming economy as it comes to be. Three steps given by state council by 2022 and 25 2030. There are listed out targets of AI economy in China. And second, the social stability. AI is going to bring fundamental transformation to the world, including China. This will lead to the issue of massive unemployment, for example. We don't need so many labors. We don't need so many people to work because AI is able to replace a lot of jobs. So what is going to happen to the social stability of China, which is Communist Party think is crucial to everything? And also AI is going to need to wider socio economic inequality. Imagine the future job requires considerable education in computer science or need a lot more education than it need now. So socio economic inequality and how this has been having a conflict with a communist society in which everybody is equal promised by the Communist Party that's challenging the ideology of Chinese Communist Party. This is going to bring fundamental challenges. And third is how I might work in the longer run with authoritarian and communist values. Some Chinese scholars has been talking about potentially whether in the longer run AI is going to help us to deliver a communist society because AI is able to provide sufficient material goods in which we don't need any human labor at all. And all the government role is to distribute those material goods. And in a world where everyone have sufficient access to material goods. Are we achieving Marxist utopia version of a communist society? That is a very interesting discussion of what the future might look like in the age of AI in the longer term.

Joanne Kuai [00:12:49]

And you mentioned that the AI could widen the inequality in the society. Can you give us some examples of that?

Jinghan Zeng [00:12:57]

I think it talks to two things. First thing is domestic socio economic inequality, and this refers you to one skill. So imagine now in the future I think we don't need taxi drivers anymore, do we? I mean, Tesla is now able to achieve a lot of self driving car in a much more safer way and people in long drives from London to Edinburgh. You don't need people there to drive. You can actually sit back, enjoy a movie or have a sleep and you will be there. So we don't need so many drivers. I think takes drivers in 1020 years will be replaced by self-driving cars. Similarly, in translation, I think I now has been playing a very important role in translation in hospital. Again, I for example, doing the scanning, doing some basic work for human, has been playing a very important role in the financial industry. I mean, replacing a lot of financial analysis is able to analyze a lot more data than a human being can do. So I think in all aspects, all the area we have already seen the impact of AI and how AI is able to do a lot of things. And I think in the longer run we don't need so many workers. So for the future job in the age of AI, if you no longer have advanced job skill, you are unlikely to be able to situated in a very good job or have a job at all. So that is going to widening the socioeconomic gap. So skill is one thing. I think skill is one thing to make need to this socio economic inequality. Second is capital. So now with more capital, if you have capital in your pockets to invest in, it's going to lead to more return to that. But if you don't have it, then you are going to the one become worse off. So the capital is also something needing to the wider socio economic inequality, and that's domestic one. You have also international socio economic inequality as well. What I have thing here is AI is really one of the probably a game changer of widening gap between developed world and developing world in the future with developed countries like United States. Europe who are able to invest in AI and to build up their strengths in those areas. Their economy is going to become stronger and stronger. But for developing countries who do not have a shortage or do not have capital and talent to invest in, the AI, zero gap with the developed country is going to become wider. So this is going to lead to international socio economic inequality as well.

Joanne Kuai [00:15:32]

So you've touched upon the race between China and us, and since we are here situated in Europe, can you tell us a little bit where does Europe sit in this global AI development and how is the AI development in China different from that of the US and Europe?

Jinghan Zeng [00:15:51]

So that's a good question. Actually, when I look into some of the debate in Europe and hearing from how European colleagues talk about AI, I think that is a question probably in minds of many policymakers in Europe. So when we talk about the AI competition, people look into America looking to China. How what is a unique approach of Europe? Some people have been characterized American way as a market oriented approach and Chinese way of state driven approach, although I do not necessarily agree in the book and you are the ethics oriented approach or customer rights oriented approach in which you are trying to put regulation first and also ensure it is an AI. What is developer ethical AI for greater goods? GDPR, for example, is one thing has been put forward by some of the voice in Europe, pointing out a unique kind of role or unique European approach. So where is Europe? I think in comparison with China and the US, the Europe do not have the same kind of resources or kind of approach with China with us. But the first approach is making your different. And how does China different from us and Europe? Well, I think different in many ways. First of all, if you compare China with the US approach, US is market oriented. China, I would say, is state driven plus market oriented. So there are two levels. One is the top level design, which the state want to achieve something at the second level, also bottom up market game there. So it's combination of a state planning plus market oriented. So I would say China is more state driven than the US and less market oriented comparison with Europe. Obviously China is also very different because China's have not been putting A6 as a key criteria driving its AI agenda. I think for China what is most important is to boosting its economy also. Now, I think in the past few years China has been increasingly pay attention to regulation and ethics of AI, but its emphasis on this matter is not comparable with that in Europe and also want to make China different from Europe and us. You will have to bear in mind China is developing country and in which it has very heavily low GDP per capita in comparison with the US and Europe. So China's low GDP per capita, low labor capital means that the market or the financial incentive for Chinese industry to use AI to replace human labour are much less than that in US and Europe, US and Europe because of the higher labour cost of the regulation, all sorts of things. Then there are much more market incentive and financial incentive to replace human labour by. But in China is is less so.

Joanne Kuai [00:19:02]

And what would implications be for the global order of AI development in China?

Jinghan Zeng [00:19:09]

Well, I think that's something my book wants to address started a bit and something I probably should make. Another paper, which I'm working on trying to do, will be more. So I think in the book I'm saying that I'm not instead of saying what might be the implication for the global order, my book are saying what might not be the implication for the global order. The book start with is how many in the US has been summarizing China's AI shortage as a geopolitical driven, top down kind of strategy to contest for global order. My book saying this is wrong because there is considerable elements of putting up considerable elements of domestic competition among Chinese actors. This would mean that what Beijing wants wouldn't necessarily be what might be implemented, and there is no coherent and clear goal of what Beijing wants to achieve. I think the. Vacation values. We shouldn't be exaggerating the implications of China's strategy for global order or what it can be. To be. Clearly, China wants to shape norms, but is still a bit far from that. And more specifically, I think this needs to next. So what if some of the implications or consequences are unintended then what they might be? I think one thing is global order is about global governance. Who shapes future of AI, Right. The relevant to what kind of norms will use, what kind of rule we use, or even what kind of knowledge AI is going to believe. And I think this is quite important. China has been working to shape global norms and trying to maintain influence in shaping the future order. But I think so far it's not very successful. If you look at the global AI regimes, you don't really find a place where China play critical roles in shaping the global order. So what China is going to do that is something very interesting. I do see a few AI regimes that are created to give China, but not a specific one, where China can exert its influence. So we're trying to do something, as China did with Asian Infrastructure Investment Bank, which China will work with bear on road country to create another set of rules for AI. That would be something to observe. And also the US China geopolitical competition has been leading to kind of securitization of making the transnational cooperation more and more difficult, putting it under the shadow of geopolitics. And this would mean transnational cooperation will be shadowed by geopolitical competition. And the ultimate consequence of that for global order is we are going to see the emerging global order. But in a very fragmented way, which would mean a very fragmented global governance structure and which are being piecemeal in different parts. And this is not necessarily a good thing because we can not coordinate a coherent and consistent human approach or global approach, global response to the challenge posed by governance.

Joanne Kuai [00:22:31]

And now you have this book recently published. Can you tell us a little bit what's your current project where some other interesting stuff you are working on?

Jinghan Zeng [00:22:40]

So in the book, actually I said up in the last chapter I set up the three major research agenda for the future, building on what I'm working on. One is how we can better understand China's strategy. I argue that it's important not only to better understand, but also its implication for others. And B is focusing on securitization in China and how this might go in the future. Third one is how also governance might evolve with AI and how it may be further embedded in that, and also how the governance part, how China's also value might be further inserted into governance, especially global governance. That's a paper I'm writing about, trying to explore how China fits in the global governance and what might be the key challenge of that. But I think I points to a relatively more pessimistic picture in which I don't find a place where China can play a role that it wants to be. I think that geopolitics is making it very difficult for China to play a major role in the future global governance. What China is going to do about that? I think that's something very interesting to be closely observed.

Joanne Kuai [00:23:57]

Yeah, I do agree. This really deserves our ongoing observation. So Jinghan, thank you so much again for joining us today. And to our listeners, you can connect with Jinghan on Twitter and with me as Joanne Kuai and I thank you for listening to the Nordic Asian podcast showcasing Nordic collaboration in studying Asia.

Outro [00:24:25]

You have been listening to the Nordic Asia podcast