Journal of Indo-Pacific Affairs, Air University Press --
This well-researched yet concise book argues that contrary to popular belief, China's artificial intelligence (AI) policies are multifaceted and are driven by economic priorities more than security priorities. Moreover, it is not a unified, top-down vision with singular focus and approach. Zeng, a Professor of China and International Studies at Lancaster University, presents strong evidence that China’s AI policies are multifaceted, but he is less persuasive when arguing they are not unified. Zeng musters policy and industry studies to support his claims but he does not disprove alternative theories and his invocation of certain academic theories of foreign policy undermine some of his arguments.
Zeng first argues that far from being a nationally concerted top-down geopolitical strategy to advance Beijing’s unified objective, China’s AI “strategy” is driven largely by economic rather than geostrategic motives. Furthermore, China’s national AI plan is an upgrade of existing local AI initiatives, reflecting bottom-up development, as opposed to national, top-down mandates. Consequently, China’s AI ‘strategy’ is more of a political manifesto than a concrete policy plan.
Secondly—though less an argument and more an observation—the Chinese central government has been “securitizing” (in the Copenhagen School sense of the word) AI to mobilize local states, market actors, intellectuals, and the general public. This security discourse reflects China’s historical anxieties about technology, regime security needs and the growing tension caused by great-power competition. Zeng warns that despite its help in convincing domestic actors, this securitization trend could undermine key AI objectives.
Lastly, Zeng presents the Chinese governance approach to the use of AI. He observes that China’s AI practices are part of a broad yet muddled adaptation strategy to governance via digital methods. AI is part of a digital technology package that the Chinese Communist Party has actively employed to not only improve public services but to also strengthen domestic control and become a global technological leader. Zeng warns that this could result in a delicate juggling act of success in delivering a booming AI economy, ensuring a smooth social transformation to the age of AI, and proving ideological superiority of its authoritarian and communist values.
Zeng offers a variety of research that supports his claim of a scattered, decentralized AI policy. It thoroughly challenges the conventional wisdom that Xi Jinping’s 2017 declarations of China’s AI goals are proof of a unified AI policy flowing down from Zhongnanhai. He cites scholarship from Ken Lieberthal and Mike Oksenberg as evidence of China’s fragmented policymaking. He also cites research from Jeff Ding (among others) as evidence of China’s AI policies being economics-oriented and often bottom-up, driven by the private sector.
While this evidence is persuasive, it accounts more for the origin of China’s AI policies, it does not fully rebut the belief that China’s current AI policies are currently national, centralized and security-focused. It may be wrong to assume China’s AI policy commenced from the top-down, but if it is being absorbed by the central government—regardless of origin—then we are still dealing with a Chinese AI policy that is more or less orchestrated by Beijing. A similar progression, arguably, describes the creation and elevation—from provincial to national—of China’s Belt and Road Initiative (BRI) policies.
Another potentially self-conflicting point is his observation that Beijing is “securitizing” AI in the Copenhagen School definition. The Copenhagen School defines “securitization” as a state actor via the utilization of a “speech act” that moves a topic away from politics into an area of security concerns, thereby legitimating extraordinary means against this socially constructed threat. Securitization of a subject is successful when the audience accepting this securitization move, undertaken by the securitizing actor. Yet the act of securitization itself in this definition is an act of a centralized, unified national government. The chapter on securitization seems shoehorned in and makes it hard to assert a decentralized, semi-chaotic process in China while referencing a centralized, nationally-controlled theory as laid out by Barry Buzan and the Copenhagen School.
Zeng asserts that Western misunderstanding of China’s AI policy inflates China as a threat. He also asserts that Beijing’s utilizing AI as a tool of securitization jeopardizes fully capitalizing on the economic benefits of AI. Zeng ends up presenting something of a contradiction: the West erroneously thinks China’s AI policy is top-down with geostrategy as a driver, yet Beijing is also “securitizing” AI with a traditional security emphasis that puts the potential economic benefits at risk. In other words, even if China’s AI policies were not originally driven by geopolitics, they increasingly will be.
Despite the tensions in perspectives, this book is concise and well-written. It offers clear definitions of different types of AI, making it intellectually accessible for laypersons. It presents research challenging certain assumptions of China’s AI strategy, which alone makes it a must-read for those interested in AI and data. It is also a must-read for scholars of the Copenhagen School seeking real-world case studies of securitization in practice. And for those who do not read it, there are certain bottom-line findings that are well-supported: China is a rising AI and data power with capabilities to challenge US interests. AI and data are and will continue to be tools to support regime stability and legitimacy in China.
Dr. Ian Forsyth
Mr. Forsyth works as a China analyst with a consulting firm supporting the US Department of Defense in Arlington, VA. He holds a PhD in international relations from the University of Southern California and a JD from Syracuse University. His views are his alone.
Disclaimer
The views and opinions expressed or implied in JIPA are those of the authors and should not be construed as carrying the official sanction of the Department of Defense, Department of the Air Force, Air Education and Training Command, Air University, or other agencies or departments of the US government or their international equivalents.