Oxford: Oxford University Press, 2024. US$35.00, cloth. ISBN 9780197682258.
Angela Zhang’s timely book on China’s big tech regulation and governance offers a valuable lens through which to examine the country’s complex domestic power structure and dynamic interactions among various actors. For many China watchers, Xi Jinping’s China is often regarded as a highly centralized system of control, particularly due to Xi’s extraordinary personalization of authority over the past decade. However, the success of policy enforcement, experimentation, and innovation often depends on local governments working in collaboration with big tech firms. In High Wire, Zhang demonstrates how the central leadership has strategically delegated regulatory authority to respond to challenges and disputes in a more effective and agile manner. Consequently, the regulatory structure governing China’s big tech sector is decentralized in practice, with local governments and technology firms playing multiple roles in both policymaking and enforcement.
Equally notable is the fragmentation of authority at the central level. For instance, prior to the establishment of the State Administration for Market Regulation (SAMR) in 2018, the responsibility for anti-monopoly regulation was dispersed among the Ministry of Commerce (MOFCOM), the National Development and Reform Commission (NDRC), and the State Administration for Industry and Commerce (SAIC) (110). Even after SAMR was established, internal bureaucratic rivalries within SAMR have continued to fragment its authority, thereby undermining regulatory effectiveness. One of the book’s major contributions to the study of China lies in its illumination of the broader characteristics of China’s political economy through the lens of big tech regulation. In doing so, Zhang develops the “dynamic pyramid” model, which is composed of three interrelated features: hierarchy, volatility, and fragility.
In particular, the logic of the feedback loop effectively captures the underlying causes of China’s sudden crackdown on big tech firms by analyzing how hierarchy, volatility, and fragility—the three elements of the dynamic pyramid model—mutually reinforce one another. Specifically, Xi’s centralized authority (hierarchy) encourages erratic regulatory enforcement (volatility), which in turn erodes market trust and predictability (fragility). This erosion of market confidence is then used to justify further party-state intervention and top-down experimentation. As a result, the governance of China’s tech sector appears to be characterized by recurring cycles of under-regulation followed by abrupt and heavy-handed crackdowns. This pattern prompts a broader question: Is China’s regulation of generative AI likely to follow a similar cycle of regulatory laxity followed by stringent intervention? While the current regulatory environment seems designed to promote AI innovation, it remains to be seen whether—and when—more assertive regulatory measures will be adopted. Identifying the conditions under which such a shift might occur remains a key question for future research.
Another issue that arises in the context of AI regulation concerns the regulatory capacity of the Cyberspace Administration of China (CAC). While the author notes that “the CAC wrestles with a scarcity of capacity and resources, in comparison to other well-established ministries such as the MIIT [Ministry of Industry and Information Technology] and the NDRC” (284), this characterization appears somewhat paradoxical. Given the CAC’s close ties to the Chinese Communist Party and its role—as addressed in earlier chapters—as a politically powerful “supra-regulator,” can it truly be said to lack regulatory capacity, even if it possesses fewer resources or less bureaucratic experience than other ministries?
Moreover, while the author argues that “the cautious stance of the United States on AI regulation is likely to have spillover effects on China, dissuading the latter from espousing an aggressive regulatory stance so as not to compromise its competitive edge in AI development” (289), I remain skeptical of this possibility—particularly in light of Xi Jinping’s 2014 assertion that “there is no national security without cybersecurity” (129). Given the CAC’s stringent control over online content and its intervention in corporate governance through the allocation of golden shares, the emergence of a cautious, deliberative regulatory approach to AI—characteristic of the US model—appears unlikely in the Chinese context.
Arguably, one of the most notable contributions of this study is its demonstration of decentralized regulation in China’s big tech sector, even under Xi Jinping’s highly centralized governance. In chapter 8, the author examines “how insecurity about state intervention has incentivized Chinese online platforms to decentralize internal governance structures in an effort to mitigate tensions arising from platform operations” (238). Alongside the active role of local courts and the control exercised by local governments over judicial personnel and resources, the self-regulation practices adopted by platform firms further illustrate how big tech regulation in China remains decentralized and fragmented, despite the overarching centralization of political authority under Xi.
Nonetheless, China’s distinctive experiment with platform decentralization does not necessarily support the author’s claim that authoritarian state regulation can foster more democratic forms of platform self-governance. Rather, such self-regulatory mechanisms seem driven less by concern for consumer interests than by firms’ efforts to safeguard their own operational security. Still, Zhang’s study lends support to the argument that Xi’s administration has emphasized an institutional and legalistic approach to system transformation—as evidenced by the upgrading of the anti-trust agency, revisions to the Anti-Monopoly Law, and the regulatory crackdown on big tech firms—all while prioritizing technological development as a national imperative. It remains to be seen whether the generative AI industry will follow a trajectory similar to that of platform firms, as outlined in the dynamic pyramid model.
Yukyung Yeo
Kyung Hee University, Seoul