美国前总统 卡特中心报告
审视科技发展对美中关系的危害
2023年2月7日
Carter Center Report Examines the Dangers of Tech Development in U.S.-China Relations
https://www.cartercenter.org/news/pr/2023/china-020723.html
近十年来，中美两国在南海、台湾等问题上的分歧越来越大，近年来分歧愈演愈烈。 美中之间的冲突正在升级，从特朗普政府在经济和技术领域的贸易战开始，然后扩展到拜登政府的人权、治理和半导体竞争领域。 随着紧张局势加剧，化解冲突的努力减少，两国陷入更大冲突的可能性增加了。
摩擦的主要来源之一与新技术的出现和使用有关，新技术不仅开辟了新的机会领域，而且还开辟了开发潜力。 今天，新技术及其应用很复杂，监管不足，而且被认为比以前的技术制度影响更深远。
两国都试图在新技术领域展开激烈竞争，这一事实使美国和中国成为战略竞争对手。
两国都有竞争优势，使他们彼此区分开来。 一方面，美国拥有资金充足的私营部门，主要为消费者进行创新。 另一方面，中国拥有国有和私营企业的国家资助机构，这些机构根据国家政策目标进行创新。 正是这种竞争优势，中国的国家驱动方式，主要涉及美国围绕个人隐私和竞争的价值观。 相反，一些中国人反感美国对技术的保护态度，在他们看来，美国的做法是试图阻止其他国家进行逆向工程或简单地使用专有技术。
由于技术的复杂性，知识产权盗窃、相互竞争的国家方法以及隐私价值观的冲突给跨国经营的美国和中国公司带来了重大挑战。 即使在有明确的技术使用法律结构的地方，对执法和对此类技术的最终控制的担忧仍然存在。 在这份报告中，我们回顾了重大新技术，讨论了美国和中国之间的利害关系问题，并描述了两国如何找到更坚实的和平共处基础的方法。 我们专门研究了人工智能、物联网、大数据和隐私、半导体和网络攻击，并提出了建设性对话的建议。
人工智能。 人工智能可以使用大数据集模拟人类的思维和反应，并且具有民用和军事应用。 基于人工智能的应用程序可用于商业目的，为客户提供最高水平的服务。 例如，该技术已被用于改善小企业的信用评分以及根据个人喜好定制产品营销。 同时，人工智能也可用于增强网络攻击的有效性或监控人员。
围绕人工智能的最大担忧之一与中美两国对人权和执法的不同看法有关。在文化上，美国人往往强调保护个人权利，而中国人则倾向于强调整个社会的权利和安全。 此外，人工智能与法律体系之间的关系在两国也不同。 例如，虽然美国和中国都使用人工智能面部识别技术来监控公民，但在如何将公民识别为国家安全威胁和提供正当程序方面存在重大差异。
除了这些差异之外，美国还批评中国通过让国家参与创新来进行不公平竞争。 [1] 值得注意的是，美国政府越来越多地参与制定创新和技术竞争政策。美国和中国都有各自政府制定的人工智能战略，但中国的战略更为全面。科技部、工业和信息化部开展这方面的研究，为产业带来新技术。此外，自2015年备受争议的《中国制造2025》发布以来，中国将人工智能发展纳入政策规划，进一步强化了中国作为美国“战略竞争对手”的地位。
由于对人权的看法和国家在创新中的作用存在根本差异，美国和中国将继续努力在这项技术上找到共同点。然而，避免未来冲突升级的能力对于在这个问题上建立信任至关重要。创造条件有助于提高透明度，并可以在发生冲突时提供缓和局势的手段，这将增加维持和平关系的机会。
物联网。 物联网或“IoT”是指人们日常使用的互连设备，包括支持技术设备的网络、数据和计算过程。 例如，互联网不再是人对人，而是“物”对“物”，通过互联网连接在一起，通过网络空间控制。 通常，设备的互连性不是由人类调节的，导致互连的网络物理系统显着增加。
物联网可用于民用和军用，并已被纳入智慧城市和智慧学校以及武器系统和飞机。 这些对象可以收集和传输大量数据，但增加连接性会带来更大的网络攻击风险。 再加上包括 5G 和人工智能在内的其他新技术，物联网带来了新的安全挑战。 黑客已经能够通过访问物联网对象（包括监控摄像头）来访问易受攻击的网络。
中国已经简化政策，将物联网融入各个行业的日常生活，包括城市建设、医药和汽车行业。
中央层面的政策推动物联网成为战略性新兴产业。
该国还制定了针对物联网设备保护隐私权的政策。
美国一直致力于防止向中国出售关键技术以保持其技术优势。 特朗普政府进一步禁止美国公司与被视为国家安全威胁的中国公司进行商业往来。 拜登总统将这一实体名单扩大到包括企业和军事研究机构，例如 SenseTime Group Limited、Leon Technology Company, Limited 和 Yitu Limited。 [2]
为了找到更坚实的基础，美国和中国都应该确保物联网必须得到足够的保护。 标准化有助于提高物联网设备的安全性，并通过提高系统之间的兼容性来确保参与全球市场。
大数据和隐私。 大数据由可能包含各种数据类型的大型数据集组成。 这些数据可用于揭示个人、公司或系统行为的复杂模式。 大数据的主要问题是它为安全和隐私创造了新的漏洞。
中国和美国都以不同的方式关注数据隐私。 美国主要关注保护可能暴露个人财务和个人信息的数据。 中国主要担心商业实体收集数据和外国间谍活动的可能性。 对数据隐私的共同关注可以为两国提供合作的起点。 这在美国尤其是一个问题，那里没有像中国的《个人信息保护法》和《数据安全法》这样包罗万象的数据隐私法。
因此，在美国改进数据保护和加强隐私标准至关重要。 如果没有全面的数据隐私法，美国将难以就此话题与中国展开富有成效的对话。
美国应弥补这一差距，并就网络安全问题与中国进行持续对话。 正如本报告所探讨的那样，高层谈判在过去被证明是成功的。 美国和中国以及其他国家之间的合作可以帮助降低数据泄露的风险。
与人工智能的情况一样，美国和中国对谁应该有权访问数据持有不同的看法。 中国政府控制在国内产生的数据，而美国个人不希望允许国家访问他们的个人数据。
两国之间的对话必须考虑到这种意识形态差异。 半导体。 半导体对于大多数现代电子产品来说都是必不可少的，并执行一系列功能，包括计算处理。 半导体行业的特点是跨国专业分工，许多步骤仅由少数几家大公司主导。 这对各个国家在陆上完整的半导体供应链造成了几乎无法克服的挑战。
近年来，美国对华为、中芯国际等需要外国技术的中国企业实施出口管制，从而减缓了中国半导体产业的发展。 美国正在努力将半导体生产转移到本土，但尽管美国半导体行业实力雄厚，但在未来许多年内，某些领域仍将严重依赖外国。
中国半导体行业在过去二十年取得了长足进步，但在半导体供应链的大部分环节仍普遍落后于全球行业领导者。 尽管高层的政策关注和国家对该行业的大力支持，中国仍然严重依赖国外设备和海外投入。
短期内，中国和美国都不太可能成功为其公司在全球半导体供应链中占据更大的份额。
虽然半导体作为一项技术的基础性质意味着合作前景有限，但两国应尽量限制对抗行动，例如扩大出口管制或惩罚性报复措施。
前进的道路。 技术包括在现有政治和社会背景下产生和运作的创新。 不同的技术意识形态方法导致了美国和中国之间的不信任。 如果不理解和接受这些根本差异，例如在私人和国家主导的创新方面，就不可能进行相互对话。
对话和外交操作程序的差异导致谈判疲劳。 加强参与需要对风险认知的相互理解，以及为技术使用制定共同规则和标准的愿望。 在两国之间建立信任将使我们更好地了解对方的意图。
我们提出了一个更有成效的双边对话的路线图。 这包括以下建议：
1. 各方都必须保持现实的期望。 管理对未来在不同问题上达成妥协的预期对于衡量成功至关重要。 任何一方都不应期望对方会彻底改变其技术竞争方式或技术在社会中的作用而进行对话。 相反，重点应该放在更好地理解不断变化的国家对潜在和新出现问题的立场上。 这有助于就基本原则形成双边共识。
2. 对话必须让合适的人聚在一起。 各种新技术的使用涉及两国的多个政府机构，因此在必要时确定与政府和私营部门的正确人员进行沟通非常重要。
三、双方应制定可行的议程。 双方都不太可能想要共享特定的知识产权，而是可能希望关注范围的两端——最广泛和最狭窄的问题。 作为进行富有成效的对话的基础，每一方都必须清楚地了解对方对合理使用网络工具规则的解释。
4. 对话必须定期进行。 这可以促使双方定期进行内部政策审查，为努力跟上快速发展的创新步伐的会议做准备。
5. 对话框架应平衡双边讨论与国际机构和协议之间的相互作用。 美国和中国不仅应该考虑双边协议和议程，而且参与者还应该考虑将这些原则和讨论扩展到更大的多边形式。
新技术不一定会破坏外交，应该用来加强而不是削弱国家之间的政治关系。
Home News & Events 2023
Carter Center Report Examines the Dangers of Tech Development in U.S.-China Relations
The latest installment in the Finding Firmer Ground report series examines how technological advancements have increased distrust and uncertainty in U.S.-China relations. From big data to semiconductors, the multiauthor report looks into the geopolitical implications of technological development and what strategies, if any, policymakers in the United States and China can take to mitigate tensions.
This report is the fourth in the Carter Center’s Finding Firmer Ground report series, which explores how dialogue, conflict management, and collaboration between the U.S. and China can sustain peace and prosperity in East Asia. Previous reports examined the role of civil society, the agricultural sector, and educational exchange in managing U.S.-China tensions.
This report is published by the Carter Center’s China Focus. The editors are Yawei Liu, senior advisor on China at The Carter Center, and Michael Cerny, program associate for the Center’s Peace Programs. You can find the report here.
The report is authored by Sara Hsu, Ph.D., clinical professor of supply chain management at the University of Tennessee at Knoxville; Ja-Ian Chong, Ph.D., associate professor of political science at the National University of Singapore and nonresident scholar with Carnegie China; Rorry Daniels, managing director of Asia Society Policy Institute; Shirley Martey Hargis, nonresident fellow in the Atlantic Council’s Global China Hub and Digital Forensic Research Lab; and John Lee, director of East West Futures Consulting.
Contact: In Atlanta, Maria Cartaya, maria.cartaya@cartercenter.org
Finding Firmer Ground: The Role of High Technology in U.S.-China Relations
https://www.cartercenter.org/resources/pdfs/peace/china/finding-firmer-ground-the-role-of-high-technology-in-u.s.-china-relations.pdf
Understanding the influence of high technology on U.S.-China relations, competition, and the prospects for cooperation.
Edited by Yawei Liu & Michael Cerny Foreword by Victor C. Shih
Foreword
II. Executive Summary
III. Introduction
IV. Artificial Intelligence
V. Internet of Things
VI. Big Data and Privacy
VII. Semiconductors
VIII. Opportunities for Dialogue
IV. Conclusion
I. Foreword
China’s accession to the World Trade Organization (WTO) marked the high-water point of bilateral relations between the U.S. and China.China was still technologically far behind the U.S. and other Organization for Economic Co-operation and Development (OECD) countries as U.S. companies looked for a large pool of cheap and adequately skilled labor.With the lowering of tariffs on Chinese-made goods after China’s WTO accession, the two countries enjoyed a decade and a half of enormous mutual gains that lowered the cost of the digital revolution in the U.S. and lifted hundreds of millions
out of poverty in China. The Chinese government also took advantage of foreign investment to absorb technologies from foreign companies and to use greater fiscal resources to build up its military and bolster industrial policy. Today, bilateral trade amounts to over $1 trillion a year, which continues to provide ample mutual benefit to firms and individuals across the two countries.
Yet, the narrowing asymmetry of power and technology have bred deeper mistrust and even hostility between the two countries. Since 2012, Chinese General Secretary Xi Jinping’s attempts to highlight China as a premier power in the world has encouraged hawks in the U.S. government to move China to the “center stage” of U.S. foreign policy with an increasing focus on China as the main competitor and strategic threat to the U.S.This report starts from a place that many from both sides can heartily agree — that the downward spiral in bilateral relations over the past six years must be managed more effectively in order to prevent worse dynamics from taking hold.
In addition, the authors, all seasoned China watchers and experts in technology policy, identify several core logics instilling distrust and overall insecurity among policymakers on both sides. They then propose potential policies to mitigate these sources of mistrust and insecurity.Fundamentally, the points raised in this report narrow down to one powerful source of insecurity, the anarchic nature of the global order, along with one strong motivation for cooperation: the profit potential of the technology industry. Taken together, these provide some hope for realistic cooperation and conflict management, which the report sets forth.
First, for years, China has taken advantage of the anarchic nature of the global order to steal billions of dollars in intellectual property (IP) from more advanced countries.
Since Chinese companies stealing technology were beyond U.S. jurisdiction, there was little that U.S.-based claimants could do to enforce American IP law on companies thousands of miles away and in other jurisdictions.The WTO provided a cumbersome mechanism for companies to lodge IP-related grievances, but it moved slowly and ultimately had no means of enforcing rulings on convicted parties (or at least no enforcement that the guilty party could not circumvent quickly). Again, this was not a problem when the technological disparity between the U.S. and China was vast. By 2010, however, Chinese technology companies like Huawei were taking market share from leading U.S. firms like Motorola and Cisco.The prospect of inadvertent or deliberate technology transfer has become a major worry among U.S. policymakers, especially in the integrated circuits (IC) sector.As the report outlines, the U.S. has rolled out a series of sanctions against the Chinese IC industry to prevent further technological leakage, going so far as to bar U.S. citizens from working for Chinese IC firms.These measures were rolled out because transnational law enforcement remains limited, so the U.S. felt compelled to maximize the reach of domestic jurisdictions to sanction individuals and companies that are domiciled in the U.S. or conduct extensive business with U.S. entities.
Even in the cybersecurity realm, the inherently anarchic nature of the global order constitutes a root problem.If the U.S. federal government wants to hack into the computer of a domestic company, it needs to obtain a court order to do so, which leaves a paper trail and potential liability. A U.S. court also can stop the government from hacking domestically, which law enforcement authorities, in theory, will enforce.
However, beyond the national border, nothing save for occasional congressional oversight stops a U.S. government agency from hacking into a foreign ompany's computer.Likewise, nothing short of a threat of retaliation stops Chinese agencies from hacking into U.S. computers.If governments around the world see retaliation as the only effective deterrent to hacking, they will engage in a constant series or hacks and counter-hacks, resulting in untold economic losses to the major powers and any countries without adequate know-how to defend themselves.
On the other hand, the authors of this report suggest that both sides have some incentive to engage in dialogue and regulatory coordination.Why would these two competing powers do that? We must recall that bilateral trade and investment has generated hundreds of billions of dollars in profits for companies on both sides. Even in the IC sector, which now is a $500 billion-a-year industry, the potential for mutual gains is enormous.
To be sure, in some areas, the relationship between the two countries is akin to a prisoner's dilemma game where cooperation by one side leads to a "sucker's” payoff and absolute gain by the other side.For example, agencies on both sides will continue to hack into the other’s government. The side that doesn’t will lose out.
Still, in other areas, coordination between the two countries will lead to gains for both sides and the lack of cooperation even by one side will lead to losses for both. In the area of renewable energy, for example, both sides have strong incentives to coordinate. Both sides will benefit from faster adoption of electric vehicles, and U.S. carmakers can benefit from the lower costs of Chinese batteries. Chinese battery makers can make money off U.S. buyers of Chinese batteries. For the U.S., increasing dependence on Chinese batteries is not such a problem since it still has plenty of oil if China were to impose a battery embargo on the U.S.
The challenge for both sides, as highlighted by the report, is to harmonize domestic or global regulations to structure incentives for coordination by both sides. Already, the U.S. has started to harmonize its state capacity in the high-tech realm so that it can match the robust technology policies of China.This should not be seen by China as an unmitigated negative development. It actually helps the U.S. government obtain better information from sectoral players about optimal policy approaches and craft intermediate policy solutions.The report makes the very sensible suggestion of forming a global regulatory body for Internet of Things (IoT) technology so that privacy standards can be coordinated and even enforced.This is an excellent suggestion. I would add that perhaps blockchain technology can be brought to bear to help with enforcement. For example, the vendor of a new IoT device must submit a blueprint of its product, which complies with a certain set of privacy requirements. The vendor must further deposit a substantial sum of money in an escrow account. If members of the governing body find a deviation in the vendor’s product from its original blueprint that jeopardizes privacy, a blockchain contract is immediately executed to take the vendor’s deposit, thus creating a financial loss for the offender.In general, the crafting of fair and transparent mechanisms with a bite will make transnational coordination in digital regulation more tractable.
In sum, competition between two sovereign nations is always tricky due to countries’inability to enforce laws on one another. However, commercial incentives, transparency, and technologies that ensure transparency provide some grounds for cooperation and conflict management in the future.
II. Executive Summary
In the past decade, China and the U.S. have found themselves increasingly in dispute over issues like the South China Sea and Taiwan, and these disagreements have only intensified in recent years. Conflict between the U.S. and China is escalating, starting with the trade war under the Trump administration on economic and technological fronts and later expanding into the realm of human rights, governance, and semiconductor competition under the Biden administration. As tensions have risen and attempts to defuse conflict have dwindled, the chance that the two countries become entangled in greater strife has increased.
One of the central sources of friction relates to the emergence and use of new technologies, which have opened not only new realms of opportunity but also potential for exploitation. Today, new technologies and their applications are complex, underregulated, and thought to be farther-reaching than previous technology regimes.
The fact that both nations are attempting to compete fiercely across the range of new technologies has situated the U.S. and China as strategic competitors.
Both nations have competitive advantages that set them apart from one another. On one hand, the U.S. has a well-funded private sector that innovates mainly for the consumer. On the other hand, China has a state-funded apparatus for both state- and privately owned enterprises that innovates based on state policy objectives. It is this competitive advantage, China’s state-driven approach, that principally concerns American values surrounding personal privacy and competition. Conversely, some Chinese resent the United States’ protective attitude toward technology, which, in their view, seeks to prevent other nations from reverse engineering or simply using proprietary technology.
Due to the complexity of technology, intellectual property theft, competing state approaches, and clashing values over privacy have created major challenges to American and Chinese companies operating cross-nationally. Even where there is a clear legal structure governing technology use, concerns remain over legal enforcement and the ultimate control of such technology. In this report, we review major new technologies, discuss the issues at stake between the U.S. and China, and describe ways that both countries can find firmer ground to coexist peacefully. We specifically examine artificial intelligence, the Internet of Things, big data and privacy, semiconductors, and cyberattacks, and conclude with suggestions for constructive dialogues.
Artificial Intelligence. Artificial intelligence can simulate human thought and reactions using large data sets and has both civilian and military applications. AI-based applications can be used for business purposes, to provide customers with the highest service levels. The technology has been used, for example, to improve credit scoring for small businesses and to customize product marketing to individual preferences. At the same time, AI can also be applied to enhance the effectiveness of cyberattacks or to monitor people.
One of the biggest concerns surrounding AI relates to different views about human rights and legal enforcement in China and the U.S. Culturally, Americans have often emphasized the protection of individual rights, while the Chinese tend to emphasize the rights and security of society as a whole. In addition, the relationship between AI and the legal system differs in both countries. While both the U.S. and China use AI facial recognition technology to monitor citizens, for example, there are key differences as to how citizens are identified as national security threats and in provision of due process.
In addition to these differences, the U.S. has criticized China for competing unfairly by involving the state in innovation.[1] Notably, the U.S. government is increasingly involved in setting innovation and technology competition policy. Both the U.S. and China have AI strategies created by their respective governments, but the Chinese strategy is more comprehensive. The Ministry of Science and Technology and the Ministry of Industry and Information Technology develop research in this area and bring new technologies to industry. In addition, China has integrated AI development into its policy plans since the controversial Made in China 2025 plan was released in 2015, furthering China’s role as “strategic competitor” to the U.S.
Due to fundamental differences in views of human rights and the role of the state in innovation, the U.S. and China will continue to struggle to find common ground on this technology. However, the ability to avoid escalation of future conflicts is essential to building trust on this issue. The establishment of conditions that help to increase transparency and can provide a means for de-escalation in the event of conflict will improve the chances of maintaining a peaceful relationship.
Internet of Things. The Internet of Things or “IoT” refers to the interconnected devices that people use daily, including the networks, data, and computational processes supporting technological devices. For example, the internet is no longer people to people, but “things” to “things” linked together through internet connectivity and controlled via cyberspace. Oftentimes, the interconnectivity of devices is not mediated by humans, resulting in a significant increase of interconnected cyber-physical systems.
IoT can be used for both civilian and military purposes and has been incorporated into smart cities and smart schools as well as into weapons systems and aircraft. These objects can collect and transmit large amounts of data, but increased connectivity comes with greater risk of cyberattacks. Coupled with other new technologies, including 5G and artificial intelligence, IoT presents new security challenges. Hackers have been able to gain access to vulnerable networks by gaining access to IoT objects, including surveillance cameras.
China has streamlined policies to integrate IoT into everyday life across a variety of industries, including the urban construction, medicine, and automotive industries.
Policies made at the central level have promoted IoT as a strategic emerging industry.
The country also has policies to protect privacy rights against IoT devices.
The U.S. has focused on preventing sale of critical technologies to China to maintain its technological superiority. The Trump administration went further to ban U.S. companies from engaging in commerce with Chinese companies that are considered a national security threat. President Biden extended this entity list to include firms and military research institutes, such as SenseTime Group Limited, Leon Technology Company, Limited, and Yitu Limited.[2]
To find firmer ground, both the U.S. and China should ensure that the IoT must be sufficiently secured. Standardization can help to increase the security of IoT devices and ensure global market participation by increasing compatibility between systems.
Big Data and Privacy. Big data is comprised of large data sets that may include a variety of data types. The data can be used to reveal complex patterns of ndividual, firm, or system behavior. The main issue with big data is that it creates new vulnerabilities for security and privacy.
Both China and the U.S. are concerned about data privacy, in different ways. The U.S. is primarily concerned with protecting data that could expose individual financial and personal information. China is primarily worried about the collection of data by commercial entities and the potential for foreign espionage. Common concerns about data privacy can provide the two countries with a starting point for cooperation. This is especially an issue in the U.S., where there are no all-encompassing data privacy laws like China’s Personal Information Protection Law and Data Security Law.
Improvements in data protection and bolstered privacy standards in the U.S. are therefore critically important. Without a comprehensive data privacy law, the U.S. will struggle to launch a productive conversation with China on this topic.
The U.S. should remedy this gap and engage in an ongoing dialogue with China on cybersecurity issues. As this report explores, high-level negotiation has proved successful in the past. Collaboration between the U.S. and China, as well as other nations, can help reduce the risk of data exposure.
As in the case of artificial intelligence, the U.S. and China hold different views on who should have access to data. The Chinese state controls data generated within the state, while American individuals do not wish to grant the state access to their personal data.
Such ideological differences must be taken into consideration in conversations between the two nations. Semiconductors. Semiconductors are essential for most modern electronics and carry out a range of functions, including computational processing. The semiconductor industry is characterized by specialized division of labor across countries, with many steps dominated by just a few major firms. This creates nearly insurmountable challenges for individual countries to onshore complete semiconductor supply chains.
In recent years, the U.S. has imposed export controls on Huawei, SMIC, and other Chinese firms that require foreign technology, thereby slowing evelopment of China's semiconductor industry. The U.S. is working to bring semiconductor production onshore, but despite the strengths of the American semiconductor industry, significant foreign dependency in certain areas will remain for many years to come.
China's semiconductor industry has made significant progress over the past two decades but generally remains well behind global industry leaders in most segments of the semiconductor supply chain. Despite top-level policy attention and significant state support for the industry, China remains heavily dependent on foreign equipment and overseas inputs.
Neither China nor the U.S. is likely to achieve success in capturing a significantly larger share of the global semiconductor supply chain for their companies over the short term.
While the foundational nature of semiconductors as a technology means that prospects for cooperation are limited, both nations should try to limit antagonistic actions, such as expansion of export controls or punitive retaliatory measures.
The way forward. Technology comprises innovations that arise and operate within an existing political and social context. Different ideological approaches to technology have given rise to mistrust between the U.S. and China. A mutual dialogue cannot occur without understanding and accepting these fundamental differences, such as with respect to private and state-led innovation.
Differences in operating procedures for dialogue and diplomacy have given rise to negotiation fatigue. Stronger engagement requires mutual understanding of risk perceptions and the desire to work toward common rules and standards for technology use. Building trust between the two nations will provide a better understanding of the other side's intentions.
We suggest a roadmap to more productive bilateral dialogue. This includes the following recommendations:
1. Each side must keep expectations realistic. Managing expectations for future compromise on the different issues is essential to measuring success. Neither side should come into dialogue anticipating that the other side will make sweeping changes in its approach to technology competition or the role of technology in society. Instead, the focus should be on better understanding the evolving state position on underlying and emerging issues. This can help to shape bilateral consensus on basic principles.
2. Dialogues must get the right people together. Use of various new technologies cuts across multiple government agencies in both countries, and it is important to identify the correct personnel to communicate with, both in government and in the private sector where necessary.
3. The two sides should set a workable agenda. Neither side is likely to want to share specific intellectual property but rather may wish to focus on the extreme ends of the spectrum — the broadest and the narrowest issues. As grounds for engaging in productive dialogue, each side must clearly understand the other side’s interpretation of rules of fair use of cyber tools.
4. The dialogue must take place on a regular schedule. This can prompt regular internal policy reviews on either side in preparation for meetings that strive to keep pace with rapidly evolving innovation.
5. The dialogue framework should balance the interplay between bilateral discussions and international institutions and agreements. Not only should the U.S. and China consider bilateral agreement and agendas, but the participants also should consider the expansion of such principles and discussions to larger multilateral formats.
New technology does not have to derail diplomacy and should be used to strengthen rather than weaken political relations between countries.
Notes:
[1] See the USTR’s Section 301 Findings on China’s technology practices,https://ustr.gov/sites/default/files/Section%20301%20FINAL.PDF
[2] Barker et al. 2022, Biden Administration Issues New Sanctions and Trade Restrictions on Chinese Technology Entities and Adds 34 Chinese Entities to the Entity List, Arnold and Porter Advisory, January
7. https://www.arnoldporter.com/en/perspectives/advisories/2022/01/biden-issues-new-sanctions-onchinese-technology
III. Introduction
As competition between the U.S. and China escalates, there are questions about how each country will use technology to further their economic and political interests. There are several issues that present potential threats to a peaceable relationship between the two nations. Although the U.S. is considered a global leader in the field of artificial intelligence and other areas of technological development, China is rapidly catching up and concerns abound that such technology will be used to accelerate cyberattacks.
At the core of the technology conflict are differences in how each country perceives
individual rights and the role of the state. The U.S. places significant weight on values
relating to individual privacy and freedom, while China views stability and state security
as paramount. Maintaining peace while advancing technological innovation requires
recognition and acceptance of these different values.
Newly developed technologies themselves present complex challenges to state values.
Artificial intelligence requires the construction of ethical guidelines that were
previously unnecessary. Access to semiconductors and the U.S.-China semiconductor
competition challenge the nature of and rules of globalization. Moreover, technologies
such as the Internet of Things and big data greatly widen the scope of technology’s
applications and create new vulnerabilities that can be exploited by third parties. These
technologies are becoming increasingly widespread across industries. These include
both civilian and military sectors, such as the health care, pharmaceutical, and
transport industries, as well as the defense industry. In China, the state has explicitly
encouraged the use of such technologies across sectors, while competition among
firms in the U.S. has rendered the adoption of new technologies a necessary tool for
firm survival.
In the following sections, we look at key new technologies and the issues they present
for the U.S.-China relationship. We elaborate on existing and potential issues regarding
new technologies. In each section, we examine how these individual technologies are
embedded in the broader framework of U.S.-China competition, describe the U.S.’ and
China’s existing policies in each technology area, and then describe overarching
practices that could help reduce the potential for conflict with regards to their
development or application . We conclude with a section about how dialogue can be
enhanced to facilitate cooperation.
IV. Artificial Intelligence
The technology and its uses. Artificial intelligence (AI) combines with big data and the
speed of cloud computing to simulate human thought and reactions. This process
reduces the costs of doing business and can improve customer service and risk
management. AI helps firms analyze patterns that may not be perceptible to humans to
uncover business risks and process transactions.
However, artificial intelligence can be used for both civilian and military purposes. For
example, AI systems can be designed to identify and exploit vulnerabilities in other
systems. AI can also be used to attack systems, such as by using human-like behavior
to crowd out legitimate users, identifying ideal targets in large data sets, and creating
vulnerabilities in consumer machine learning models to distort results.
Overall, the U.S. possesses high levels of domestic and international AI talent and welldeveloped AI innovation and hardware systems. China has many AI patents and high
venture capital investments, but fewer AI companies and lower levels of AI talent than
the U.S. A major reason for this difference is that many Chinese AI researchers move to
the U.S. to work for cutting-edge firms and institutions. While both countries are
leaders in the number of AI companies, China is unique in developing specialized zones
called New Generation AI Innovation Development Experimental Zones, where AI
demonstrations and policy tests are carried out.
In addition, many AI-targeted microchips are developed by American companies, on
which China is partially dependent. The U.S. dominates production of graphics
processing units produced by American firms like Nvidia, Google’s Tensor Processing
Unit (TPU), and field programmable gate arrays (FPGAs) from companies like Intel and
Xilinx.
U.S. companies also have an innovation edge in AI-enabled analytical tools. For
example, the U.S. leads the world in claiming the top deep learning frameworks,
TensorFlow developed by Google, and PyTorch, developed by Facebook.[1] The U.S. is
also a leader in natural language processing, with over double the number of firms as
China and triple the number of employees.[2] Moreover, the U.S. is a global leader in
computer vision and autonomous driving.
Seeking to catch up to the U.S., China has used its large population and market to its
advantage in recent years. China has made strides in obtaining data for AI training
purposes because it has been able to take advantage of customer data and customer
biometrics. Before the Personal Information Protection Law came into effect in October
2021, privacy laws were more lenient in China, which provided Chinese firms and
government entities with rich databases on which to train artificial intelligence models.
China also leads in speech recognition.
In addition, China has tasked selected private sector AI “champions” with specific
innovations, such as Alibaba for smart cities and Baidu for autonomous vehicles. This
status comes with enhanced access to state finance and preferential treatment in
contract procurement.[3] Alibaba implemented the City Brain project across China to
increase the speed and transparency of administrative processes and improve service
delivery, such as in Hangzhou's Xiaoshan International Airport, where Alibaba
technology was used to better manage airport schedules and maintenance.[4] Notably,
some of the AI champions, including iFlyTek, SenseTime, and Huawei, have been
banned in the U.S. on national security grounds and are not permitted to acquire
hardware, software, and technology from American companies.
Some artificial intelligence software relies on visual recognition, and China’s prowess in
this area is undeniable. In 2016 and 2017, for example, Chinese scientists won first
place at the Large Scale Visual Recognition Challenge for computer vision systems.
Causing international outcry, China has deployed its advanced technology in this area
to identify and trace individuals in China’s Xinjiang Uyghur Autonomous Region (XUAR)
who the state views as potential terrorists, often using race-based attributes. By using
SenseTime’s technology to compare faces stored in a database of known terrorist
suspects with real-time surveillance footage, China’s security apparatus is now able to
identify and locate these suspects. Megvii’s “Sharp Eyes” project accomplishes the
same task across multiple provinces, including and beyond Xinjiang. Its cameras
collaborate with the Integrated Platform for Joint Operations so that police officers can
carry out facial identification at a wider scale.
This reflects a difference in ethical norms for technology use between the U.S. and
China, as the Chinese government views the broad use of surveillance and facial
recognition as tools for maintaining stability and security, whereas U.S. firms have
restricted its use among the public to some extent on the basis of human rights and
individual privacy. Sullivan (2021) argues that AI’s surveillance capacity represents “the
most significant single obstacle to cooperation on AI norms, rules, and procedures
between our nations.”[5] It should be noted, however, that U.S. agencies are using facial
recognition in criminal investigations where criminal activity has been committed,
and are planning to increase use of such software[6], even if facial recognition for the
purposes of rating citizens and identifying suspicious behavior of noncriminals is not
used as it is in China.
The GAO report on facial recognition states that 15 of the 42 federal agencies surveyed
on the use of facial recognition software use nonfederal sources, with only one of these
agencies keeping records of what nonfederal software is being used. Six of the
agencies reported using the technology to identify suspects engaged in civil unrest,
riots, or protests. By contrast, companies such as Facebook have eschewed the use of
facial recognition due to concerns raised by privacy experts.
Natural language processing has allowed the Chinese government to identify
individuals for national defense purposes. The company iFlytek uses natural language
processing to assist the Ministry of Public Security in creating a national speechpattern and voice database.[6] The National Security Agency in the U.S. also uses
natural language processing to convert communications to text. In addition, the U.S.
Defense Department’s Defense Advanced Research Projects Agency (DARPA) uses the
Deep Exploration and Filtering of Text program to extract and act upon information
across unstructured text data for the purposes of defense.[7] Executive Order 12333 on
United States Intelligence Activities calls upon departments and agencies to provide
“the President, the National Security Council, and the Homeland Security Council with
the necessary information on which to base decisions concerning the development and
conduct of foreign, defense, and economic policies, and the protection of United States
national interests from foreign security threats.”[8] This order has permitted the U.S.
government to monitor citizens without the use of a warrant, and it is unclear to what
extent electronic surveillance is used.
Although both China and the U.S. are using AI to monitor citizens, albeit to varying
extents, the underlying difference in attitudes toward human rights provides a basis for
conflict between the two nations. What is at issue is less the use of technology to
identify criminals than differences in what is outlawed. China has no protections in
practice for freedom of speech or assembly and therefore classifies a wide range of
government critics as national security threats. These include individuals who speak
out against the state, such as lawyers, feminists, religious practitioners, and in many
cases those belonging to ethnic minority groups. The U.S. also targets individuals who
are considered national security threats under a narrower definition while generally
upholding due process. The difference in attitudes toward human freedoms such as
freedom of speech and religion, protection against unreasonable search and seizure,
and due process has resulted in fear that the Chinese government will apply its internal
standards of privacy to foreign users of its equipment and software. This fear is at the
root of the conflict over AI monitoring and surveillance.
Compounding this fear is the fact that China has implemented AI in a more integrated
fashion between its civil and military spheres, as opposed to the U.S., which faces a
divided audience in AI implementation between the Pentagon and Silicon Valley.[9] As a
result, Chinese tech firms have built in back doors or security vulnerabilities, possibly
for the purposes of providing information about domestic Chinese users to the Chinese
intelligence community. For example, a Chinese camera vendor, Xiongmai, was found to
have created software containing an undocumented back door that could access
millions of cameras. While some of the code may be prepared for China’s domestic
market, it is challenging to maintain separate software installations for devices sent to
other destinations abroad. In other words, technology shipped from China may still
contain this vulnerability[10], which can present problems for other states’ national
security.
There is also anxiety in the U.S. that China will use AI to counter traditional American
defense-related espionage. The U.S. National Security Commission on Artificial
Intelligence released its final report in March 2021,[11] which stated that China is a
strategic competitor to the United States in AI. It further indicated that the U.S. is
worried that China will use AI to identify and expose American sources and methods.
In addition, the U.S. views China’s development of AI as a threat due to China’s
cyberattack targets. For example, China’s hack of Microsoft Exchange is believed to be
part of an unknown long-term artificial intelligence project.[12] The hack of email data
was carried out by Hafnium, a In addition, the U.S. views China’s development of AI as a
threat due to China’s cyberattack targets. For example, China’s hack of Microsoft
Exchange is believed to be part of an unknown long-term artificial intelligence project.
[13] The hack of email data was carried out by Hafnium, a cyber-espionage group with
alleged ties to the Chinese government. Cyberattacks can be made more rapidly, with
better precision, and in greater secrecy with the use of AI. Cyberattacks have been
used in recent years to steal trade and government secrets. There is a concern that use
of AI-fused data for blackmail, deepfakes, or swarms are possible in the future. China
has also allegedly used deepfakes to influence Taiwanese elections.[14]
China’s fear of U.S. cyberattacks has likewise grown. Chinese media reports have
stated that the U.S. is the largest source of cyberattacks in the country, attacking
aerospace, scientific research institutions, large internet companies, and government
agencies.[15] At the same time, there appears to be less fear in China that the U.S. will
harness AI to augment cyberattack capabilities.
Finally, the U.S. believes that China can use AI to offset U.S. military superiority by
implementing a type of “intelligentized war” that uses alternative logistics,
procurement, training, and warfare algorithms.
In In this scenario, battle networks will connect systems, and armed drones with
autonomous functions will be employed. AI can also help to identify and hit valuable
targets more rapidly. Notably, China has expressed the wish to ban usage, though not
development, of autonomous lethal weapons.[16]
The Department of Defense’s 2020 annual report to Congress on the Chinese military
stated that China views new technologies as the means to engage in intelligentized
warfare. This will speed up decision-making capabilities and improve intelligence and
surveillance technologies. The People’s Liberation Army even reorganized its research
and education institutes to integrate new technologies with the development of new
operational concepts in 2017.
The U.S. military also uses AI and possesses autonomous weapons. The U.S. has made
use of the technology under Project Maven, which was tasked with identifying
insurgent targets in Iraq and Syria. AI research is being conducted in the areas of
intelligence collection and analysis, logistics, cyber operations, information operations,
command and control, and semiautonomous and autonomous vehicles.[17] However, it
is critical to note that the U.S. has a transparent policy on AI decision-making in
offensive capabilities; it has published guidance on Autonomy in Weapon Systems,
which requires “appropriate levels of human judgment over the use of force.”[18]
As in the case of AI-based surveillance and data usage, U.S. anxiety over China’s
“intelligentized war” concept is based on China’s record of behavior on other
international norms and values. For example, Morgan et al. (2020) claim that China has
not complied with biological and chemical weapons treaties, nor with its World Trade
Organization obligations, leading to concerns over China’s commitment to binding
agreements.[19] In addition, China has proposed a ban on lethal autonomous weapon
systems that defines such systems quite narrowly, and which is not supported by the
US. This has led some observers to call into question China’s true commitment to
banning such weapons, especially as U.S. standards are currently stricter than those in
China’s proposed ban.
Still, the U.S. itself ignores international agreements when they fail to serve its
interests. This has been a feature of American participation in the international arena
since before its independence from Great Britain. More recently, the U.S. has withdrawn
from the Paris Climate Accords and has violated WTO tenets under the Trump trade
war. Although the Biden administration has sought to reinvest in international
institutions and agreements, there are no guarantees that future administrations will
maintain compliance and few punishments the international system can impose on its
largest power for violating the rules.
Current policies. Both the U.S. and China have implemented policies to promote and
safeguard new technologies. AI has received ample consideration by both governments.
The Biden administration released a blueprint for a “bill of rights” in October 2022 to
guide the design and implementation of technologies such as artificial intelligence. The
objective was to ensure safe systems, protect individual privacy, and prevent
algorithmic discrimination.[20]
Prior to that, on Feb. 11, 2019, President Trump signed Executive Order 13859,
establishing the American Artificial Intelligence Initiative coordinated by the National
Science and Technology Council Select Committee on Artificial Intelligence. This
program coordinates AI activities across all federal departments and streamlines
objectives related to improving access to federal data, models, and computing
resources, lowering barriers to AI technology implementation, reducing vulnerability to
attacks, training American AI researchers, and implementing a plan to safeguard U.S.
economic and national security interests. The American AI Initiative sought to create a
blueprint for setting up the National AI Research Resource. On March 19, 2019, the
government released AI.gov, which provides the public with information on federal
government activities in AI. The National AI Initiative Act of 2020 was released to
coordinate federal government efforts to accelerate AI research and application.[21]
The U.S. National Artificial Intelligence Research and Development Strategic Plan,
updated June 2019, aims to meet eight priorities in support of the AI Initiative: to make
long-term investments in AI research, to develop methods for human-AI collaboration,
to address ethical and legal aspects of AI, to ensure safety and security of AI systems,
to develop shared data sets for AI training, to create benchmarks in AI, to understand AI
R&D workforce needs, and to increase public-private partnerships for development of
AI.[22] The Biden administration has set up this National Artificial Intelligence
Research Resource Task Force to understand how to promote innovation.
The Department of Defense has published a classified AI strategy and carries out
related tasks, such as setting up a Joint Artificial Intelligence Center to implement
artificial intelligence technologies, publishing a roadmap for AI development, and
creating a National Security Commission on Artificial Intelligence to assess militaryrelated AI technologies and make recommendations for further implementation.[23]
The report recommends that the U.S. achieve military AI readiness by 2025 through
Pentagon leadership reforms and augmentation of the Department of Defense’s AI R&D
portfolio.[24]
In China, the government plays a particularly important role in AI research and
development. Both the Ministry of Science and Technology and the Ministry of Industry
and Information Technology coordinate and develop research in this area. In addition,
the Ministry of Industry and Information Technology plays an important role in bringing
new AI technologies to industry. The NDRC’s Department of High-Tech Industry also
plays a key role in promoting technological advancement.
Like Biden’s AI “bill of rights,” China has released several regulations aimed to protect
society. These include rules for online algorithms, certification of “trustworthy AI
systems,” and establishment of AI principles.[25]
China has a number of plans and policies that have integrated AI into state goals. The
Made in China 2025 plan was released in May 2015 and aims to develop intelligent
products and production. AI was also highlighted in the 13th Five-Year Plan, the Robot
Industry Development Plan, the 13th Five-Year Plan for National Technological
Innovation, the Special Campaign on Advancing Innovative Development of Intelligent
Hardware Industry, and the 13th Five-Year Plan on Developing Emerging Sectors of
Strategic Importance (Deloitte 2019).
China’s State Council issued the New Generation Artificial Intelligence Development
Plan in July 2017, which comprises much of China’s AI strategy. The plan aims to
accelerate innovation in the area to make China’s artificial intelligence field
internationally competitive by 2020 and to become a world leader by 2030. At the end
of 2017, China detailed some of these goals in the Three-Year Action Plan on Promoting
the Development of New Generation AI Industry. Local governments across the country
have also issued their own policies detailing how they will enhance AI development.
China’s 2019 defense white paper, “China’s National Defense in the New Era,” states
that new technologies can increasingly be applied to the military sphere and that
military-civil fusion can work toward modernizing China’s military forces.[26]
Proposals to find firmer ground. One of the biggest issues in the U.S.-China
relationship regarding artificial intelligence is whether AI-based applications can be
used to automate lethal actions. While China has expressed the desire to ban the use of
automated lethal weapons, the U.S. has refused to negotiate a new international
agreement on autonomous weapons.[27] In addition, neither country has committed to
ending the development of such technologies, since they view maintaining artificial
intelligence capabilities as essential to winning future wars.
However, China has taken steps toward creating ethical standards in AI. China’s “White
Paper on Artificial Intelligence Standardization” published by the Standards
Administration of China outlines three principles for the ethical use of AI technologies.
[28] These are the principle of human interest, which asserts that AI should benefit
human welfare; the principle of liability, which states that there should be
accountability for the development and deployment of AI-infused technology systems;
and the principle of consistency in rights and responsibilities for commercial entities to
protect their intellectual property. Other entities have created their own AI ethics
standards, including the Beijing Zhiyuan Artificial Intelligence Research Institute, which
established the Artificial Intelligence Ethics and Safety Research Center, putting
forward the “Artificial Intelligence Beijing Consensus."[29] In July 2021, the Ministry of
Science and Technology laid out the “Ethical Norms for New Generation Artificial
Intelligence,” which states that AI technologies should respect human rights and
privacy.[30]
In the US, ethical standards on AI have been adopted by the Department of Defense.[31]
These standards include: requiring responsible development and deployment of AI
capabilities among Department of Defense personnel; minimizing unintended bias due
to AI capabilities; requiring transparency of AI methodologies and data sources;
ensuring reliable AI capabilities that are safe, secure, and effective; and requiring AI
capabilities to be governable, preventing unintended consequences and permitting
deactivation of deployed systems where necessary.
The U.S. Intelligence Community has also laid out ethics regarding AI. These provide
guidance on how to develop and use AI. Ethical standards include respecting the law
and protecting civil rights; ensuring AI method and use transparency and
accountability; reducing bias; applying human judgment in cases where an action may
infringe upon civil liberties; ensuring security and resilience using best practices; and
using AI that has been informed by science and technology.
As noted above, however, there are some ethical differences between the Chinese and
American approaches to AI use. The difference can be reduced to varying views of
human rights and legal enforcement/commitment. On one hand, the United States is
primarily concerned with individual rights as codified in the Constitution’s Bill of Rights,
while China is more focused on the rights and security of the citizenry as a whole as led
by the Chinese Communist Party. For example, Americans may view freedom of speech
as inviolable, while the Chinese may see this freedom as a threat to social stability.
These ethical differences will certainly impact the use of AI in both the civilian and
military arenas and must be addressed before they create larger rifts between the two
countries.
Perhaps, in this case, the U.S. can recognize China’s right to use AI in domestic affairs
while restricting China from using the technology in the United States. This suggestion
is highly unlikely, however, since the U.S. has already crossed this line by sanctioning
China for its treatment of the Uyghur people within China’s borders and banning
imported products from Xinjiang. China views many ethnic Uyghurs as terrorists and
violates their human rights by using vast AI systems to identify and track alleged
terrorists. This suggests that the U.S. needs to ask itself additional questions about
limits to its ability to enforce American human rights standards in other countries. After
the U.S. has crossed China’s so-called “bottom line,” how far is it able and willing to go
to enforce its perspective?
However, neither the Chinese nor the American commitment to international
agreements is fully credible. Both nations are likely to breach, reject, and refuse to sign
or ratify such agreements if they conflict with domestic interests. International
agreements are extremely difficult to enforce. As Koplow (2013) notes, the world’s
leading international judicial tribunal, the International Court of Justice, does not have
jurisdictional power over the United States or China because neither has submitted
itself to the court’s authority.[32] The United Nations Security Council also holds the
power to resolve disputes, but the veto power held by each of its five permanent
members, including China and the U.S., guarantees that these members will be
protected from adverse findings.
Monitoring systems to understand when AI is used maliciously is also essential. The
Defense Advanced Research Projects Agency (DARPA) in the U.S. can detect and
disrupt malicious information campaigns.[33]
An eventual goal should be to avoid unintended escalation, but that may be difficult to
achieve at the outset given high levels of distrust. An interim goal, given that anything
concrete is difficult amid the political climate in both the U.S. and China, could be to
establish conditions that open the way for discussions about mechanisms for
transparency, confidence-building, and de-escalation. This could be in terms of
monitoring, limiting the deployment of particularly escalatory AI technologies, and
perhaps the use of hotlines so long as humans are part of the decision-making process.
The lead up to Cold War-era arms control talks between Russia and the United States
could prove instructive given the early lack of transparency and concerns with
escalation, even if it is an imperfect analogy. Nonetheless, everything remains tentative
until the political climate is more accepting of forward movement, meaning that
informal contact and exploration may be all that can be hoped for at the moment.
Notes:
[1] Laskai, Lorand and Helen Toner. 2019. Can China Grow Its Own AI Tech Base? Chapter in STANFORDNEW AMERICA DIGICHINA PROJECT AI POLICY AND CHINA Realities of State-Led Development, Special
Report No. 1 October 29, 2019, Edited by Graham Webster
[2] Deloitte. 2019. Global artificial intelligence industry white paper,
https://www2.deloitte.com/content/dam/Deloitte/cn/Documents/technology-mediatelecommunications/deloitte-cn-tmt-ai-report-en-190927.pdf.
[3] Roberts, Huw, Josh Cowls, Jessica Morley, Mariarosaria Taddeo, Vincent Wang, Luciano Floridi. 2020.
The Chinese approach to artificial intelligence: an analysis of policy, ethics, and regulation, AI and
Society (2021) 36:59–77.
[4] See page: https://www.alibabacloud.com/blog/city-brain-now-in-23-cities-in-asia_595479
[5] Sullivan, Ryan. 2021. The U.S., China, and Artificial Intelligence Competition Factors. China Aerospace
Studies Institute, https://www.airuniversity.af.edu/Portals/10/CASI/documents/Research/Cyber/2021-10-
04%20US%20China%20AI%20Competition%20Factors.pdf?
ver=KBcxNomlMXM86FnIuuvNEw%3D%3D.
[6] GAO. 2021b. Facial Recognition Technology: Federal Law Enforcement Agencies Should Better
Assess Privacy and Other Risks. GAO, June 3, https://www.gao.gov/products/gao-21-518.
[7] Sullivan, Ryan. 2021. The U.S., China, and Artificial Intelligence Competition Factors. China Aerospace
Studies Institute, https://www.airuniversity.af.edu/Portals/10/CASI/documents/Research/Cyber/2021-10-
04%20US%20China%20AI%20Competition%20Factors.pdf?
ver=KBcxNomlMXM86FnIuuvNEw%3D%3D.
[8] Eggers, William D., Neha Malik, and Matt Gracie. 2019. Using AI to unleash the power of unstructured
government data. Deloitte Insights, January 19,
https://www2.deloitte.com/us/en/insights/focus/cognitive-technologies/natural-language-processingexamples-in-government-data.html.
[9] Department of Defense. 2008. Executive Order 12333 United States Intelligence Activities,
Department of Defense, https://dpcld.defense.gov/Portals/49/Documents/Civil/eo-12333-2008.pdf.
[10] Horowitz, Michael C., Gregory C. Allen, Elsa B. Kania, and Paul Scharre. 2018. Strategic Competition
in an Era of Artificial Intelligence. CNAS Paper, https://www.indexinvestor.com/resources/ResearchMaterials/NatSec/Strategic_Competition_in_Era_of_AI.pdf.
[11] Medin, Milo and Gilman Louie. The 5G Ecosystem: Risks & Opportunities for DoD Defense Innovation
Board, 3 April 2019.
https://media.defense.gov/2019/Apr/03/2002109302/-1/-1/0/DIB_5G_STUDY_04.03.19.PDF
[12] National Security Commission on Artificial Intelligence. 2021. Final Report, National Security
Commission on Artificial Intelligence, https://www.nscai.gov/wp-content/uploads/2021/03/Full-ReportDigital-1.pdf
[13] Temple-Raston, Diana. 2021. China's Microsoft Hack May Have Had A Bigger Purpose Than Just
Spying, NPR, August 26. https://www.npr.org/2021/08/26/1013501080/chinas-microsoft-hack-may-havehad-a-bigger-purpose-than-just-spying
[14] Taipei Times. 2021. Task force at work to combat Chinese deepfake videos. Taipei Times, November
4, https://www.taipeitimes.com/News/taiwan/archives/2021/11/04/2003767291.
[15] Xinhua. 2021. Zhao Lijian said the United States is the world's largest source of cyber attacks,
Xinhua, July 20, http://www.xinhuanet.com/world/2021-07/20/c_1127674897.htm
[16] Kania, Elsa. 2018. China’s Strategic Ambiguity and Shifting Approach to Lethal Autonomous
Weapons Systems. Lawfare Blog, April 17, https://www.lawfareblog.com/chinas-strategic-ambiguityand-shifting-approach-lethal-autonomous-weapons-systems.
[17] Congressional Research Service. 2020. Artificial Intelligence and National Security Updated
November 10, 2020 , Congressional Research Service Report, https://fas.org/sgp/crs/natsec/R45178.pdf.
[18] Forrest E. Morgan, Benjamin Boudreaux, Andrew J. Lohn, Mark Ashby, Christian Curriden, Kelly Klima,
Derek Grossman. 2020. Military Applications of Artificial Intelligence: Ethical Concerns in an Uncertain
World, RAND Research Publication, https://www.rand.org/pubs/research_reports/RR3139-1.html.
[19] Forrest E. Morgan, Benjamin Boudreaux, Andrew J. Lohn, Mark Ashby, Christian Curriden, Kelly Klima,
Derek Grossman. 2020. Military Applications of Artificial Intelligence: Ethical Concerns in an Uncertain
World, RAND Research Publication, https://www.rand.org/pubs/research_reports/RR3139-1.html.
[20] White House. 2022. “FACT SHEET: Biden-Harris Administration Announces Key Actions to Advance
Tech Accountability and Protect the Rights of the American Public,” October, White House,
https://www.whitehouse.gov/ostp/news-updates/2022/10/04/fact-sheet-biden-harris-administrationannounces-key-actions-to-advance-tech-accountability-and-protect-the-rights-of-the-american-public/
[21] AI.gov. 2022. AI.gov National AI Initiative Act. AI.gov, accessed September 28,
https://www.ai.gov/#:~:text=The%20National%20AI%20Initiative%20Act,economic%20prosperity%20an
d%20national%20security.
[22] National Science and Technology Council. 2019. The National Artificial Intelligence Research and
Development Strategic Plan: 2019 Update. A Report by the Select Committee on Artificial Intelligence of
the National Science and Technology Council, June, https://www.nitrd.gov/pubs/National-AI-RDStrategy-2019.pdf.
[23] Congressional Research Service. 2020. Artificial Intelligence and National Security Updated
November 10, 2020 , Congressional Research Service Report, https://fas.org/sgp/crs/natsec/R45178.pdf.
[24] National Security Commission on Artificial Intelligence. 2021. Final Report, National Security
Commission on Artificial Intelligence, https://www.nscai.gov/wp-content/uploads/2021/03/Full-ReportDigital-1.pdf
[25] Sheehan, Matt. 2022. China’s New AI Governance Initiatives Shouldn’t Be Ignored, January 4,
Carnegie Endowment, https://carnegieendowment.org/2022/01/04/china-s-new-ai-governanceinitiatives-shouldn-t-be-ignored-pub-86127
[26] Uber, Maj. Richard. 2020. China’s Artificial Intelligence Ecosystem. NIU Research Monograph,
https://ni-u.edu/wp/wp-content/uploads/2021/08/Uber_Monograph_DNI2021_02261.pdf.
[27] Wareham, Mary. 2018. Campaign to Stop Killer Robots: Report on Activities, Convention on
Conventional Weapons Group of Governmental Experts meeting on lethal autonomous weapons systems
United Nations, Geneva, 9-13 April 2018.
[28] Ding Jeffrey, Triolo Paul (2018) Translation: excerpts from China’s “White Paper on Artificial
Intelligence Standardization.” New America. https://www.newamerica.org/cybersecurity-initiative/digic
hina/blog/translation-excerpts-chinas-white-paper-artificial-intelligence-standardization/.
[29] Sun, Mingchun. 2021. Ethical norms and industry self-discipline in the development of artificial
intelligence technology (in Chinese), Sina news, January 14, https://finance.sina.com.cn/tech/2021-01-
14/doc-ikftpnnx6825029.shtml.
[30] Georgetown University Center for Security and New Technology. 2021. Ethical Norms for New
Generation Artificial Intelligence Released, Georgetown University Center for Security and New
Technology. October 21, https://cset.georgetown.edu/publication/ethical-norms-for-new-generationartificial-intelligence-released/
[31] Department of Defense 2020. DOD Adopts Ethical Principles for Artificial Intelligence, Department of
Defense website, February 24,
https://www.defense.gov/Newsroom/Releases/Release/Article/2091996/dod-adopts-ethical-principlesfor-artificial-intelligence/.
[32] David A. Koplow. 2013. Indisputable Violations: What Happens When the United States
Unambiguously Breaches a Treaty, Fletcher Forum of World Affairs 37(1): 53-74.
[33] National Security Commission on Artificial Intelligence. 2021. Final Report, National Security
Commission on Artificial Intelligence, https://www.nscai.gov/wp-content/uploads/2021/03/Full-ReportDigital-1.pdf
V. Internet of Things
The technology and its uses. The Internet of Things refers to the network of billions of
objects worldwide that are connected to the internet. These objects contain chips and
often sensors that send data back to databases for analysis. They include wearable
devices, auto parts, household appliances, and medical devices.
The Internet of Things can be used for both civilian and military purposes. IoT is
increasingly incorporated into the construction of smart cities, which can improve
residential services, transportation, and public utilities. Smart schools can help track
attendance and meal payments, while utilities can take advantage of smart lighting and
smart meters. Intelligent transportation systems and smart parking can help increase
public transportation efficiency. Even so, looking at individual cities, there were no U.S.
or Chinese cities among the top 10 cities of the IMD Smart Cities Index for 2021.[1]
Additionally, IoT is used for military purposes to make up for workforce challenges.
Creating a warfighting network makes for speedier intelligence collection and threat
identification. This process includes collecting data through sensors on numerous
platforms, such as weapon systems, aircraft, and troops. The IoT-connected sensors
and radars collect and transmit data on the positioning and movement of U.S. troops
and countries of interest.[2]
Application of IoT lags in other areas in the U.S., including within the federal
government. The U.S. Government Accountability Office found that while many federal
agencies use IoT to monitor equipment, control facility access, or track physical assets,
agencies that did not intend to use IoT viewed such devices as having low returns.[3]
While some agencies use IoT, their application mainly focuses on specific objects rather
than on an IoT ecosystem and, more significantly, automated decision-making or data
analysis.
On balance, China assesses that the IoT has a higher value in streamlining public
services than the U.S. China is building up urban IoT infrastructure, as directed by
several central government departments. This infrastructure seeks to accelerate IoT
infrastructure by 2023, setting up IoT demonstration bases, digital villages, smart
transportation, smart construction, and smart agriculture by 2023.[4] This method is far
ahead of U.S. government implementation or planning.
The NATO Science and Technology Committee report on the Internet of Things finds
that this technology has great potential in military applications[5]. IoT devices can be
used in sensors for command, control, communications, computers, intelligence,
surveillance, and reconnaissance (C4ISR) systems to gather and transmit data.
Additionally, the IoT can be used for firepower control systems to respond
automatically to threats. Mobile technologies applied to IoT devices, such as
smartphones, can provide access to tracking or mapping applications. IoT devices can
also track shipments and log movements within logistics systems.
However, there are many vulnerabilities associated with the IoT. The increased
connectivity of IoT devices may lead to congestion across networks, which could block
the functionality of critical devices, such as medical devices. Malicious network attacks
can also lead to device malfunctions, which could be costly to human life or critical
infrastructure. Russia’s invasion of Ukraine showcases that the IoT can be a vehicle for
information warfare. Against the backdrop of financial sanctions from the U.S., Canada,
and the European Union, hackers have employed cyberattacks to combat Russia.
Anonymous, a hacker group, has declared a cyberwar on Russia by hacking stream
services.[6]
IoT also increases the potential cyber-attack and cyber-accident surface, rendering
new vulnerabilities where there were fewer or different vulnerabilities before. The
power, telecom, and IT industries may be critical as attacking these industries would
have significant negative consequences. With the growth of 5G and artificial
intelligence, there is an increased possibility of attacks on IoT devices. Both
technologies speed up the rate at which attacks could occur, increasing the likelihood
of success within any given timeframe. 6G looms large as China has already begun
working on the infrastructure, while the U.S. is still overcoming security challenges
from an unstable 5G rollout.
The concern for the U.S. is that the global power that controls 6G will command the rest
of the century. 6G would be used in military operations and our day-to-day lives.
Currently, China enjoys unprecedented influence on the global stage regarding the
diffusion and deployment of advanced communications technologies. 6G would have an
air latency of less than 100 microseconds and is expected to be 100 times faster than
5G, with broader network coverage and enhanced reliability. With the implementation
of 6G-based solutions, IoT will continue to become increasingly integrated into people's
lives and connect 10 times more devices per square kilometer and significantly more
connected devices to come.[7]
Concerning military applications, the benefits of IoT that beneficially transform modern
warfare are also causing concern about malicious cyberattacks. The tremendous
impact of IoT is the increase of available methods and opportunities for data gathering,
yet the trouble is that hackers can also take advantage. Concerns include vehicle
safety, healthcare, and supply chains. Hackers could commandeer vehicles, take
control of medical devices, and disrupt supply chain operations. In addition, IoT devices
may provide hackers with critical information. In March 2021, hackers gained access to
Silicon Valley's Verkada Inc., infiltrating companies’, hospitals’, prisons’, schools’, and
police departments' live feeds that included 150,000 surveillance cameras. Hackers
viewed videos from psychiatric hospitals and Verkada's offices.[8]
There is also a concern in the U.S. that China will set international standards for IoT
devices. This trend would allow China to use IoT more effectively for surveillance and
collecting intelligence. A report prepared for the U.S.-China Economic and Security
Review Commission stated that “the combination of widespread adoption of IoT
products and Chinese research into exploits raises the threat of unauthorized access to
U.S.-based IoT devices and the networks they connect to.”[9] The authors assert that
IoT devices manufactured in China are targets for exploitation.
Another issue underscored in the U.S.-China Economic and Security Review
Commission report is that the IoT contains many different devices and systems with
multiple providers of endpoints, gateways, and networks. The systems require
compatible standards so that product designers and consumers will be able to
purchase and use devices interchangeably.
To some extent, the issue between the U.S. and China on IoT stems from the fact that
China has had streamlined policies to integrate IoT into everyday life across a variety of
industries. China is ahead in this area, which gives the country an edge in using the
technology to serve its own purposes. China uses the IoT in smart cities for real-time
data collection, in industry to optimize operations, in medicine to improve patient care
and collect medical data, and in smart cars to sense how vehicles relate to the road
environment. As a result, the U.S. would benefit from policies that furthered technology
in this area as well so that it can remain competitive.
Current policies. In the U.S., recent policy has increased IoT security at the federal
level. Recently, Congress approved the Strengthening American Cybersecurity Act of
2022. This act covers critical infrastructure and the federal government. It includes
mandatory cyber incident reporting by owners of critical infrastructure within a
specified time frame of a cyber breach. Despite the passing of this legislation, cyber
incident reporting is often complicated because of the incentives behind choosing not
27
to report and the differences between what is reported to the government and to the
public.[10] Existing reporting regulations are too narrow and insufficiently
standardized. For example, in 2015, the Office of Personnel Management was hacked,
negatively impacting about 22 million personnel records. This breach took place in June
but was not reported until April due to a lack of reporting regulations and standards.
The 2022 act’s cyber incident report requires reporting from critical infrastructure
owners four days from the breach. Timely reporting on cyber incidents assists with
faster responses.
The IoT Cybersecurity Improvement Act also became law at the end of 2020. This act
required the National Institute of Standards and Technology (NIST) and the Office of
Management and Budget to raise cybersecurity for the IoT devices among federal
agencies, including creating and producing standards and guidelines for the
government on best practices to ensure that principles and policies are consistent with
NIST’s standards and guidelines.[11]
Chinese focus on IoT began in 2009 with the inclusion of the sector in Premier Wen
Jiabao’s work report as one of five “strategic emerging industries.” The State Council’s
2010 decision on strategic new emerging industries also promoted IoT. MIIT and the
National Development and Reform Commission have laid out tasks and priorities for IoT
development. State and local departments have rolled out policies promoting IoT
development.[12]
With regard to IoT, the U.S. has prevented sale of critical technologies to China to
maintain its technological prowess. The Obama administration in 2015 prevented Intel,
Nvidia, and Advanced Micro Devices from selling highly sophisticated supercomputer
chips to China to avert their use in military devices. Two years later, in 2017, the Trump
administration barred the sale of the Lattice Semiconductor, which holds
programmable software that provides an alternative way to build AI chips, to a Chinesebacked investor.[13] In 2022, the Biden administration banned advanced semiconductor
technology exports to China.
The Committee on Foreign Investment in the United States (CFIUS), which
recommended that the sale of Lattice be blocked, helps to protect strategically
important technologies. CFIUS’ ability to review foreign investments was expanded
under the Foreign Investment Risk Review Modernization Act of 2018.
In addition, in May 2019, President Donald Trump issued an executive order that banned
U.S. companies from using information and communications technology from
companies considered a national security threat. As Huawei’s products were judged to
28
be insecure, it was added to the entity list by the Commerce Department the same day.
In June, the U.S. added five other companies, including Chengdu Haiguang Integrated
Circuit, Chengdu Haiguang Microelectronics Technology, Higon, Sugon, and Wuxi
Jiangnan Institute of Computing Technology.
President Joe Biden has extended this policy and included more firms on the entity list.
These companies include Aviation Industry Corporation of China, Proven Glory Capital,
and Proven Honour Capital, which have served as financial arms of Huawei and sold
bonds to international investors.[14] The Commerce Department also added a
substantial number of Chinese military-related research institutes and companies to
the entity list, including the Academy of Military Medical Sciences and 11 research
institutes in response to emerging technologies for "brain-control" weapons. Chinese
military-civil fusion, or MCF, is the Chinese government's strategy to reach its goal to
create the most technologically advanced military in the world. Entities that produce
biotechnologies supporting China's MCF strategy were also placed on the entity list.
Furthermore, in response to China's human rights violations in Xinjiang, the Biden
administration added a substantial number of related entities to a U.S. investment
blacklist. These entities include SenseTime, China's top artificial intelligence firm,
identified as developing facial recognition programs to track Uyghurs.[15]
Proposals to find firmer ground. The U.S. and China both need to recognize that the
Internet of Things remains vulnerable to exploitation and needs to be further secured.
Rules and standards in both countries should increase the security of such devices, and
network-scanning software should be updated to notify owners of intrusions. Consumer
privacy rules must be applied to IoT devices to limit sensitive data collected, and IoT
devices should disclose potential data exposure issues.
China already has some regulations to protect privacy rights against IoT devices. The
Measures for the Protection of Information Security Levels contain five security levels
for information and data management systems. IoT devices that collect customer data
fall under this regulation.
Standardization, in general, can also help to ensure that neither the U.S. nor China is
locked out of participation in the other country’s markets due to incompatible systems.
International standardization bodies can help to bring together national standardssetting departments. Within the U.S. government, a department needs to be
responsible for developing IoT or 5G standards within America. By contrast, China has
expanded its standardization work by focusing on modernization and standardization of
industry through its 2018 revised Standardization Law and China Standards 2035
project.
29
Notes:
[1] IMD. 2021. Smart City Index 2021. IMD publication, https://www.imd.org/smart-city-observatory/home/.
[2] Lockheed Martin. 2017. IOT is Transforming Modern Warfare, Lockheed Martin website,
https://www.lockheedmartin.com/en-us/news/features/2017/internet-of-things-transofrming-modernwarfare.html
[3] GAO. 2020. Internet of Things: Information on Use by Federal Agencies. GAO Report GAO-20-577,
https://www.gao.gov/assets/gao-20-577.pdf.
[4] Ministry of Industry and Information Technology. 2021. Notice on Printing and Distributing the "Threeyear Action Plan for the Construction of New Internet of Things Infrastructure (2021-2023)." Ministry of
Industry and Information Technology Notice [2021] No. 130, http://www.gov.cn/zhengce/zhengceku/2021-
09/29/content_5640204.htm.
[5] Tonin, Matej. 2017. The Internet of Things: Promises and Perils of a Disruptive Technology. NATO
Science and Technology Committee Report, https://www.nato-pa.int/download-file?
filename=/sites/default/files/2017-11/2017%20-%20175%20STCTTS%2017%20E%20bis%20-
%20INTERNET%20OF%20THINGS.pdf.
[6] Dan Milmo. 2022. Anonymous: the hacker collective that has declared cyberwar on Russia, The
Guardian, February 27, https://www.theguardian.com/world/2022/feb/27/anonymous-the-hackercollective-that-has-declared-cyberwar-on-russia
[7] Qadir et al. 2022. Towards 6G Internet of Things: Recent advances, use cases, and open challenges.
ICT Express, https://www.sciencedirect.com/science/article/pii/S2405959522000959
[8] Horowitz, Michael C., Gregory C. Allen, Elsa B. Kania, and Paul Scharre. 2018. Strategic Competition in
an Era of Artificial Intelligence. CNAS Paper, https://www.indexinvestor.com/resources/ResearchMaterials/NatSec/Strategic_Competition_in_Era_of_AI.pdf.
[9] John Chen, Emily Walz, Brian Lafferty, Joe McReynolds, Kieran Green, Jonathan Ray, and James
Mulvenon. 2018. China’s Internet of Things, Research Report Prepared on Behalf of the U.S.-China
Economic and Security Review Commission October 2018,
https://www.uscc.gov/sites/default/files/Research/SOSi_China's%20Internet%20of%20Things.pdf
[10] Mary Brooks, Sofia Lesmes. 2022. Last Call at the “Star Wars Bar”: Harmonizing Incident and Breach
Reporting Requirements, Lawfare Blog, July 5, https://www.lawfareblog.com/last-call-star-wars-barharmonizing-incident-and-breach-reporting-requirements
[11] Scott Ikeda. 2020. IoT Cybersecurity Improvement Act Signed Into Law: New Security Requirements
for Federal Government Devices, CPO Magazine, December 18, https://www.cpomagazine.com/cybersecurity/iot-cybersecurity-improvement-act-signed-into-law-new-security-requirements-for-federalgovernment-devices/
[12] Lee, John. 2021. The Connection of Everything: China and the Internet of Things, MERICS report, June
24, https://merics.org/en/report/connection-everything-china-and-internet-things.
30
[13] Wang You and Chen Dingding. 2018. Rising Sino-U.S. Competition in Artificial Intelligence, China
Quarterly of International Strategic Studies, Vol. 4, No. 2, 241–258
[14] Jennifer Jacobs. 2021. Biden blocks 59 Chinese companies including Huawei in amended Trump order,
Business Standard, June 4. https://www.business-standard.com/article/international/biden-blocks-59-
chinese-companies-including-huawei-in-amended-trump-order-121060400185_1.html
[15] Jeanne Whalen. 2021. U.S. bans investment in Chinese surveillance company SenseTime, saying it
supports repression of Uyghur minority population, Washington Post, December 10,
https://www.washingtonpost.com/technology/2021/12/10/us-investment-ban-sensetime/
31
VI. Big Data and Privacy
The technology and its uses. Big data refers to massive data sets that can be analyzed
to reveal otherwise undiscovered patterns and associations. Big data tends to be
generated at high volumes per second across a large variety of data types. Such large
amounts of data can then be used to better determine customer preferences, supply
chains, or market risks. Larger amounts of data, particularly coupled with faster
processing speed and technologies such as artificial intelligence and cloud computing,
result in better understanding of patterns that indicate everything from customer
creditworthiness or market trends to more efficient design of roads and logistic chains.
Big data may be subject to attack in various ways. For example, the Hadoop framework,
which is a commonly used platform for big data, has known vulnerabilities that must be
closed to reduce the risk and damage of cyberattacks. The framework was initially built
without strong security considerations, and security was patched on in later versions.
Hadoop’s weak points include potential for password leakage to particular
applications, user or group information storage corruption, denial of service issues, user
account impersonation, and more. As many institutions use Hadoop, this creates
serious concerns over data security. The threat is compounded by the fact that lists of
companies using Hadoop are available over the internet. Other big data frameworks
face similar issues.
Data privacy is an issue that was amplified during the Trump years, particularly among
technology companies that collect and process big data. The Trump administration
attempted to ban Chinese social media company Tiktok in the United States, as well as
to restrict operations of WeChat and Alipay over fears of Chinese data access.
According to Samm Sacks (2020), the issue of data privacy between the United States
and China is complex in part because the U.S. lacks a comprehensive data privacy
regulation that addresses the issues.[1]
Among the targets for hacking in the U.S. have been large troves of personal data,
which may reveal exploitable individual vulnerabilities. These include the exfiltration of
personal data on federal employees from the U.S. Office of Personnel Management in
2015 in addition to hacks on health insurance firms, tax preparation companies, and
hotel chains that provide a huge trove of data.[2] Access to such data, which can then
be analyzed using artificial intelligence, can enable actors to target key personnel and
compromise them by identifying personal weaknesses.[3] One reason such
vulnerabilities exist is inadequate protections on big data in the United States, a
problem that persists despite scandals over the collection and use of big data from
Facebook by Cambridge Analytica.[4]
32
China, too, worries about data privacy, albeit relating to the collection and storage of
data by commercial entities. For example, a Reuters report from November 2021 states
that the Cyberspace Administration of China sought a delisting of the ride-hailing app
company Didi from the U.S. stock market due to data security concerns.[5] This move
may have to do with the significant amounts of personal data such apps store,
potentially including information from state sources. A November 2021 report by Now
from Hong Kong revealed that a dating app was able to retrieve personal information
about individuals from the Public Security Bureau.[6] A concern for the Chinese state is
addressing what they see as the potential for foreign espionage, cyberattacks, and
malign collection and manipulation of data.
Both the United States and China have entities that collect large amounts of individual
social media data, which can affect personal privacy and even subject individuals to
harassment. Such efforts tend to be state-related in China and associated with profitmaking enterprises and elections in the United States, though state-related efforts in
China do also have commercial tie-ups and applications. Recent investigative reports
point to China-related entities collecting large amounts of social media information
relating to policymakers, academics, and others of interest to the state.[7] Chinese
security agencies are also allegedly collecting information on social media accounts to
silence critics online, including to intimidate family members.[8] With regard to the U.S.,
Cambridge Analytica represents the most prominent example of data collected on
individual social media accounts in support of disinformation relating to election
campaigns.[9] Whistleblower accounts detail Facebook and other social media
companies extracting user data to drive advertising profits, even to the detriment of
users.[10] Security flaws further complicate risks from such activities.
Even if the Chinese state collects and manages large amounts of data on individuals, it
is more ambivalent about giving commercial entities greater access, as seen in the
passage and implementation of Personal Information Protection and Data Security laws
in 2021.[11] Under the Data Security Law, data from China must be stored locally rather
than in overseas servers consistent with its insistence on cyber sovereignty. Like many
Chinese laws, these new pieces of legislation give the state significant remit. Violations
can result in hefty fines and blacklisting for corporations.[12] That said, these laws are
consistent with rising public concerns about data privacy in relation to commercial
interests among Chinese citizens (Liu, 2020).
Common concerns about big data and data privacy in the hands of commercial entities
may, therefore, provide some grounds for U.S.-China cooperation. Despite a lack of
coordinated regulation and resistance by social media firms, there appears to be a
growing momentum in the United States for more regulation on the data that
33
corporations can collect and store in the wake of the latest scandal facing
Facebook/Meta.[3] There may be some basis for Washington and Beijing to come up
with common standards of regulating commercial use of big data given shared
skepticism toward the power that such information provides to corporations. Limiting
what commercial firms collect and store can also reduce security risks for both the
United States and China. Washington and Beijing can further work with the European
Union, which had earlier put forward their General Data Protection Regulation (GDPR)
to begin addressing some of these issues, including the right to erasure.[14]
Current policies. China’s cybersecurity policies are governed by the Cybersecurity Law
of the People's Republic of China, which establishes obligations of internet service
providers, personal information protection, and information infrastructure security.
China also has other regulations, including the Information Security Technology —
Implementation Guide for Classified Protection of Information System and the
Information Security Technology — Classification Guide for Classified Protection of
Cybersecurity. China’s data regime is protected by the Cybersecurity Law, which came
into effect in June 2017. The Cybersecurity Law aims to protect citizens and
organizations in the cybersecurity realm. An underlying consideration of the Chinese
approach toward online information and privacy is the concept of cyber sovereignty,
where the state should have control of all information generated and used within its
jurisdiction, including citizens and corporations abroad (Creemers, 2020).
China’s Personal Information Protection Law (PIPL) provides a degree of online privacy
protection vis-à-vis commercial entities. PIPL limits the data that corporations can
collect on individuals, including consent, access, rectification, and erasure, while
restricting the transfer of individual data outside of China’s borders. This includes
corporate human resources data such as employees’ compensation and performance
data. Like Europe’s GDPR, PIPL provides extraterritorial jurisdiction to the Chinese
state, affecting not only corporations with a physical presence in China but also those
that conduct business with entities located in China. Unlike the GDPR, however, PIPL
neither restricts nor provides oversight of state action. PIPL also has no mandates on
the safe destruction of data.
The related Data Security Law regulates the processing and transfer of data overseas.
It does so by categorizing different levels of data. Most strictly controlled is “core” data
pertaining to national interests, followed by “important” data that includes information
relating to national interests and individuals. Any transfer of such data overseas,
including the handing over of data to foreign law enforcement agencies and judiciaries,
requires differing levels of official approval. Generally, any personal or consumer data
collected in China must be stored within China and not transferred abroad. Like PIPL,
the Data Security Law has extraterritorial reach, again meaning that commercial
34
entities that do business with individuals and businesses in China are subject to these
regulations even if they do not have a physical or legal presence within China.
The U.S. does not have an all-encompassing data protection law, relying on a
combination of federal and state laws to protect privacy. The Federal Trade
Commission Act allows the U.S. Federal Trade Commission to bring actions against
firms that engage in unfair or deceptive data privacy activities or that fail to provide
sufficient security of personal information. Federal sector-specific laws also strive to
protect personal data privacy. State laws may restrict use of personal data as well.
Congressional testimony over Facebook/Meta’s use of user data is currently fueling
discussion on a need for legislation to further regulate the collection and use of
personal data available on social media.
Data privacy is an issue that is gaining greater public attention in the United States, but
there remains no comprehensive legislative or policy approach parallel to China’s PIPL
and Data Security Law or the European Union’s GDPR. Public awareness of the risks
surrounding data privacy increased dramatically with the scandal surrounding the sale
and use of Facebook data by the firm Cambridge Analytica to influence both the United
Kingdom’s Brexit referendum and the 2016 U.S. presidential election. The matter gained
further public traction with the growth in popularity of Chinese-owned social media
service TikTok and leaks by Facebook/Meta whistleblower Frances Haugen.
Nonetheless, legislative and policy responses are piecemeal at present, opening the
United States to espionage risks and U.S. citizens to cybercrime as well as excessive
corporate manipulation of their personal data.
Common concerns in both Washington and Beijing about the collection and exploitation
of user data may provide some basis for cooperation between the two sides.
Washington likely has reason to protect users from harassment and risks from state
actors like China and Russia as well as exploitation and misuse by technology firms like
Facebook/Meta and TikTok/ByteDance. Beijing has a desire to prevent external state
and corporate actors from accessing personal data about its citizens and companies
under its concept of cyber sovereignty. This confluence of interests provides some
grounds for discussion, confidence-building, and even coordination between
Washington and Beijing over regulation of the collection, sharing, and storage of
personal and corporate information on social media and other platforms. Even if the
United States and China initially embark on unilateral efforts to regulate these areas,
their convergence of interests on such issues may provide future opportunities for
seeking understanding.
Cyberattacks. In an environment of increasing political tension, the potential for future
U.S.-China cyberconflict is on the rise. Some experts have commented that the
35
U.S.-China new technology rivalry looks like a “digital cold war.”[15] However, at
present, much of the conflict between the two nations appears to be in the area of
politics and trade, and cyberconflict has not been a major focus.
The U.S.-China technology rivalry has been characterized as a type of cold war because
there is no “opting out” of new technologies for either country. If either the U.S. or China
spends much of its time maintaining mature systems without investing in new systems,
that country will lose any technology and security advantage it might have had.
Utilization of new technology is essential, as are funding for research and development
and policies to promote implementation.
As new technologies are increasingly implemented, the potential for cyberconflict will
inevitably grow. One reason for this is that new technologies create new vulnerabilities.
AI and 5G in particular will greatly expand the number of feasible cyberattack surfaces.
Another reason is that new technologies can be adapted for malicious purposes.
To some extent, China is less vulnerable than the U.S. because China has a limited
number of ports through which the domestic internet is connected to international
networks. This means that China could shield itself more readily from large-scale
cyberattacks. China’s media is also heavily controlled, increasing the possibility for
third parties to spot content-related security hacks as well as reducing citizen
sensitivity to government intervention in new technologies.
Both nations have cybersecurity units as part of national defense departments. For
example, the U.S. set up a Cyber Command in 2009 to combat cyberattacks and has
used numerous cyberwarfare tactics in physical conflicts. China’s Third Department of
the People’s Liberation Army consists of cybersecurity forces, and the People’s
Liberation Army includes computer network operations as part of its military
operations.
The U.S. and China have so far not been engaged in a cyberwar per se, but the two
countries have been involved in cyberconflict. Cyberwar can be defined as attack and
defense on and of computer systems, including hardware and/or software, while
cyberconflict may constitute smaller acts of aggression that do not escalate to the
scale and intensity of a cyberwar.
Both sides have participated in cyberconflict, however. The U.S. allegedly originates
most offensive cyberattacks on China. A report by China’s National Computer Network
Emergency Response Technical Team (CNERTT) found that there was a 91% increase in
cyberattacks by the U.S. on China in 2018, infecting 3,607 Chinese websites.[16] The
U.S. often seeks information related to military and government organizations through
cyberattacks, and the U.S. National Security Agency regularly spies on Chinese
computers and networks.
36
That said, CNERTT does not appear to clearly distinguish between state-related and
non-state cyberattacks on China that may originate from the United States.
Under the Obama administration, Chinese hacking of the U.S. declined markedly due to
a bilateral agreement in 2015 to stop hacking intended to steal trade secrets. Hacking
picked up again under the hawkish Trump administration, which issued a report by the
U.S. Trade Representative in March 2018 detailing cases of Chinese cybertheft against
the United States.[17] At the end of 2018, China stepped up cyberattacks on U.S. critical
infrastructure in the areas of energy, financial, transportation, and healthcare (Finkle
and Bing 2018). Most of the cyberattacks, however, were focused on stealing
technological secrets. In response to Chinese cyberattacks, the U.S. has been carrying
out counter-cyberattacks against Chinese intelligence and military targets. In addition,
Zhu Hua and Zhang Shilong, two Chinese nationals, were accused of participating in
hacking campaigns that targeted several U.S. government agencies, including the
Energy Department, laboratories at NASA, and the U.S. Navy. International Business
Machines Corp. and Hewlett Packard Enterprise Co. are among companies whose
computer-services operations were breached by hackers, who then used that access to
burrow into their clients’ networks. China’s hacking campaign allegedly aimed to target
technology services providers that support businesses with technologies such as cloud
storage.
In 2021, a major hack of Microsoft Exchange servers that U.S. officials and experts
attributed to China may represent a return to more aggressive cyberattacks, but the
aim of the attack may be to build up a database of personal information of Americans
rather than just commercial.[18] In 2022, it was reported by Chinese media that the U.S.
National Security Agency infiltrated China’s telecommunications networks. The Global
Times, a state media outlet, asserted that the U.S. stole key technology data, such as
network equipment configuration, network management data, and operational data.[19]
Recent U.S. policy has improved the ability of the U.S. to respond to cyberattacks. The
Presidential Policy Directive 41 of July 2016 dictates a federal response to cyberattacks
to either the public or private sectors. A May 2017 executive order, “Strengthening the
Cybersecurity of Federal Networks and Critical Infrastructure,” states that the
executive branch has the authority to control cybersecurity risk for critical
infrastructure. The order notes that the government will respond rapidly to attacks in
collaboration with the private sector. In addition, in 2018, U.S. President Trump reversed
President Obama’s Presidential Policy Directive 20 under the classified National
Security Presidential Memorandum 13, which allows the U.S. government to use
powerful cyber weapons.
The stakes of becoming a cyberattack victim are high for both nations, especially
37
regarding critical infrastructure. Therefore implementation of cybersecurity must be
viewed as equally important to implementation of new technologies at the micro-level.
In addition, it is essential that new technology firms and industries gain protection from
external use and exploitation. In some cases, only the government can block
transactions that could threaten pioneering new technology industries.
Despite the rising stakes of cyberconflict and the conflict-ridden political environment,
the best bet either country has at reducing the possibility of conflict is to work
together. Improved relations can deter the regular use of damaging cyberattacks
between the U.S. and China by increasing the costs associated with such behavior.
We recommend that the U.S. set up a coordinated effort to enhance cybersecurity by
protecting networks, databases, the Internet of Things, and other critical technology
infrastructure. We also recommend that the U.S. set up an ongoing dialogue with China
to address cybersecurity issues between the two countries that also covers data
collection, storage, and use. Big data, after all, is becoming an increasingly common
target for cyberattacks and can be exploited to compromise both commercial and
security interests. Finally, we recommend the creation of a new global governance
institution that can address cyberconflict and help to ensure cyber peace.
The diplomatic component to maintaining peace in cyberspace cannot be stressed
enough. High-level negotiation has been shown to be effective in deterring Chinese
hackers from attacking U.S. targets under the Obama administration and, even more
importantly, has played a major role in building up U.S.-Chinese political and economic
relations over the past 40 years. The U.S. and China should make better use of the HighLevel Joint Dialogue on Cybercrime and Related Issues and the Law Enforcement and
Cybersecurity Dialogue, engaging military leaders in the process as well.
Proposals for finding firmer ground. The U.S. needs to implement data rules for all
firms, both domestic and foreign, without blocking data flows to the U.S. Regulations
must take into account national security and privacy concerns, considering the extent
to which data collected is sensitive or risky and how the data is used. A federal data
privacy law would address these issues as well as cross-border data flows and
collection and storage of personal information by foreign firms. This would reduce the
focus on China as a strategic competitor and create a more systematic and rational
means of treating data usage. Without a clearer sense of standards on the U.S. side,
trying to move forward on trying to establish some sort of understanding with other
actors, including China, would be highly challenging.
The U.S. and China, along with the European Union, share interests in the regulation of
big data and privacy protection. Publics in these jurisdictions demand limits on how and
38
to what extent their personal data can be used for commercial purposes beyond their
original intention. Such commonality provides some basis for cooperation or at least
coordination on regulating the types of data that commercial entities can collect, store,
transfer, use, and sell, as well as establishing protocols for consent and the safe
destruction of commercially held data. There is likely to be resistance in the United
States from technology firms whose business models rely heavily on harvesting, using,
and selling personal data. However, mobilizing public support to overcome lobbying
efforts should be possible.
Collaboration on these matters by the United States, China, Europe, and others can
afford the public better protection of privacy while limiting the exposure of states and
corporations to espionage risks. Even if such regulation restricts what any individual
state can exploit regarding adversaries, competitors, and rivals, they can take away an
element of contestation and reduce friction. Governments have an interest to move
forward on better privacy protections regarding big data, at least in the commercial
realm. Indeed, this is what the European Union and China have done with recent
legislation. The United States should follow suit, especially given that models from which
to take reference now exist, such as the European Union’s GDPR.
There should also be thinking about how to relate discussions about big data and privacy
to cybersecurity, which are two distinct but linked issues. Ongoing dialogue with China to
address cybersecurity issues between the two countries should also cover data
collection, storage, and use. Big data, after all, is becoming an increasingly common
target for cyberattacks and can be exploited to compromise both commercial and
security interests. There should be some coverage of cybercrime as well, given the
persistent risk of states working with cybercriminals to engage in cyberattacks.
However, there needs to be recognition that the United States and Europe view data and
privacy in a way that is fundamentally different from China. Like Russia, China holds a
commitment to cyber sovereignty.[20] All information generated within the state belongs
to the state. This can extend to its citizens and companies operating overseas. The U.S.
and Europe tend to believe in a need to keep information from the prying eyes of the
state, even if this sometimes means not scrutinizing corporations sufficiently.
Such basic philosophical differences in thinking undergird legal frameworks in the
United States and China and need to be accounted for in any discussion on data and
privacy. A possible way forward may be to focus first on areas of agreement — such as
the need to protect data — as a basis to build trust before moving on to more difficult
topics. Coordination on this front can be technical and less potentially contentious since
both Washington and Beijing have an enduring interest in protecting their public as well
as private data.
39
Notes:
[1] Samm Sacks. 2020. Data Security and U.S.-China Tech Entanglement. Lawfare Blog, April 2,
https://www.lawfareblog.com/data-security-and-U.S.-China-tech-entanglement.
[2] Adams, Michael. 2016. Why the OPM Hack is Far Worse than You Imagine. Lawfare Blog, March 11,
https://www.lawfareblog.com/why-opm-hack-far-worse-you-imagine; Graff, Garrett M. 2020. China’s
Hacking Spree will Have a Decades-Long Fallout. Wired. November 2,
https://www.wired.com/story/china-equifax-anthem-marriott-opm-hacks-data/; Peterson, Andrea. 2015.
2015 is Already the Year of the Healthcare Hack--and It’s Only Going to Get Worse. Washington Post,
March 20, https://www.washingtonpost.com/news/the-switch/wp/2015/03/20/2015-is-already-the-yearof-the-health-care-hack-and-its-only-going-to-get-worse/
[3] Temple-Raston, Diana. 2021. China's Microsoft Hack May Have Had a Bigger Purpose Than Just
Spying, NPR, August 26. https://www.npr.org/2021/08/26/1013501080/chinas-microsoft-hack-may-havehad-a-bigger-purpose-than-just-spying
[4] Fruhlinger, Josh. 2020. The OPM Hack Explained: Bad Security Practices Meet China’s Captain
America. CSO Online, February 12, https://www.csoonline.com/article/3318238/the-opm-hack-explainedbad-security-practices-meet-chinas-captain-america.html; Lapowski, Issie. 2019. How Cambridge
Analytica Sparked the Great Privacy Awakening. Wired, March 17,
https://www.wired.com/story/cambridge-analytica-facebook-privacy-awakening/
[5] Zhu, Julie, Kane Wu, and Brenda Goh. 2021. Beijing Presses Didi to Delist from U.S. Over Data Security
Fears -- Sources. Reuters. November 26, https://www.reuters.com/world/china/china-asks-didi-delist-ussecurity-fears-bloomberg-news-2021-11-26/
[6] Now新聞. 2021. 「交友網站經過?數據配對 連接公安系統認證?份」. 《Now新聞》, November 19,
https://news.now.com/home/local/player?newsId=457163
[7] Cate Cadill. 2021. China Harvests Masses of Data on Western Targets, Documents Show. Washington
Post, December 31,
https://www.washingtonpost.com/national-security/china-harvests-masses-of-data-on-western-targetsdocuments-show/2021/12/31/3981ce9c-538e-11ec-8927-c396fa861a71_story.html
[8] Xiao, Muyi and Paul Mozur. 2021. A Digital Manhunt: How Chinese Police Track Critics on Twitter and
Facebook. The New York Times, December 31, https://www.nytimes.com/2021/12/31/business/chinainternet-police-twitter.html?referringSource=articleShare
[9] Hern, Alex. 2018. Cambridge Analytica: How Did It Turn Clicks into Votes? The Guardian, May 6,
https://www.theguardian.com/news/2018/may/06/cambridge-analytica-how-turn-clicks-into-voteschristopher-wylie
[10] Morris, Loveday, Elizabeth Dwoskin, and Hamza Shaban. 2021. Whistleblower Testimony and
Facebook Papers Trigger Lawmakers Call for Regulation. Washington Post, October 25,
https://www.washingtonpost.com/technology/2021/10/25/facebook-papers-live-updates/
[11] Horwitz, Josh. 2021. China Passes New Personal Data Privacy Law, to Take Effect Nov. 1. Reuters,
August 20, https://www.reuters.com/world/china/china-passes-new-personal-data-privacy-law-takeeffect-nov-1-2021-08-20/
40
[12] Burgess, Matt. 2021. Ignore China’s New Data Privacy Law at Your Peril. Wired, November 5,
https://www.wired.com/story/china-personal-data-law-pipl/
[13] Dan Milmo, 2021. Facebook and Instagram gathering browsing data from under-18s, study says, The
Guardian, Nov 16, https://www.theguardian.com/technology/2021/nov/16/facebook-and-instagramgathering-browsing-data-from-under-18s-study-says
[14] European Union. 2021. What is GDPR, the EU’s New Data Protection Law? GDPR.eu.
https://gdpr.eu/what-is-gdpr/.
[15]Marc Champion. Digital Cold War, 2019. Bloomberg Quick Take, December 12,
https://www.bloomberg.com/quicktake/how-u-s-china-tech-rivalry-looks-like-a-digital-cold-war
[16] Lindsey, Nicole. 2019. New CNCERT Report Shows Most Cyber Attacks on China Originate from
United States, CPO Magazine, June 24, https://www.cpomagazine.com/cyber-security/new-cncertreport-shows-most-cyber-attacks-on-china-originate-from-united-states/
[17] USTR. 2021. U.S.-E.U. Trade and Technology Council (TTC), Accessed December 8, 2021,
https://ustr.gov/useuttc
[18] Temple-Raston, Diana. 2021. China's Microsoft Hack May Have Had a Bigger Purpose Than Just
Spying, NPR, August 26. https://www.npr.org/2021/08/26/1013501080/chinas-microsoft-hack-may-havehad-a-bigger-purpose-than-just-spying
[19] Kharpal, Arjun. 2022. “Chinese state media claims U.S. NSA infiltrated country’s telecommunications
networks,” CNBC, September 22, https://www.cnbc.com/2022/09/22/us-nsa-hacked-chinastelecommunications-networks-state-media-claims.html
[20] Creemers, Rogier. 2020. China’s Approach to Cyber Sovereignty, Konrad Adenauer Stiftung, Berlin,
Germany,
https://www.kas.de/documents/252038/7995358/China’s+Approach+to+Cyber+Sovereignty.pdf/2c6916
a6-164c-fb0c-4e29-f933f472ac3f?version=1.0&t=1606143361537
41
VII. Semiconductors
The technology and its uses. Semiconductors are a foundational technology for
virtually all modern electronics. They can be divided into broad categories of
application, among which there are different types of integrated circuits (ICs).
Examples of ICs used throughout society include processors that provide the “brains”
for computers, memory chips that store information, and signal converters that
interface between digital and analog signals (for example, converting digital
information into sound waves in electronic audio devices). Other types of
semiconductors are also becoming increasingly common and in demand, as technology
becomes increasingly digitized. Notably, semiconductors are used in sensors, which are
increasingly ubiquitous as the Internet of Things expands.
The imperative to squeeze more computing power from smaller devices has led to a
phenomenon described by “Moore’s Law,” which observes that the number of
transistors on an IC doubles about every two years. Today, even a pocket calculator has
immensely more processing power than the computer that guided Apollo 11 to the moon
in 1969. Pushing forward this technological frontier has required progressively greater
technical know-how and larger investments, with the returns to successful firms rising
in tandem. The result has been global market consolidation in the semiconductor
manufacturing business: Over the past two decades, the number of companies
operating leading-edge semiconductor fabrication plants has fallen from around 20 to
just two, namely South Korea’s Samsung and Taiwan’s TSMC.
Similar pressures have operated on other steps of the semiconductor value chain.[1] The
result is a semiconductor industry that is structured by a highly specialized division of
labor between different countries, with many niches dominated by a handful of
companies. Prominent examples include the TSMC-Samsung duopoly in leading-edge
fabrication (manufacturing of the physical chips), and the monopoly in extreme
ultraviolet lithography (EUV) systems, which are required for cutting-edge fabrication,
by the Netherlands firm ASML. The complexity of the technologies involved and the
incumbent advantages make these industry leaders effectively unchallengeable within
their niches over the short term.
These supply chain characteristics run against the political imperative to “onshore”
semiconductor production, which is being inflated by the flow-on effects of
semiconductor shortages during the COVID-19 pandemic and intensifying international
competition in strategic technologies. The U.S. has “weaponized” the semiconductor
supply chain through export control measures targeting Huawei, SMIC, and other
42
Chinese firms that rely on foreign inputs to perform their core business. In response,
China is redoubling its efforts to build domestic industries’ capacities along all
segments of the semiconductor supply chain to mitigate risk from political tensions
with the U.S. and its allies (Lee and Kleinhans, 2021). The European Union, South Korea,
Japan, and Taiwan are also running programs aimed at bringing a bigger chunk of the
semiconductor supply chain within their borders, framed by the language of
“technological sovereignty,” “strategic autonomy,” and supply chain security.
In addition to being a ubiquitous foundational technology, semiconductors are also a
critical enabler for emerging technologies such as artificial intelligence (AI). The U.S.
National Security Commission on AI in its final report of 2021 recommended restricting
the export of certain semiconductor-related technologies to China in order to curb
development of Chinese AI-enabled military capabilities.[2] Rising computerization of
devices is making semiconductors critical to a range of economically important sectors
such as the automotive industry, where the effects of semiconductor shortages have
focused U.S. and European government attention on the implications of foreign
dependence.
As the Internet of Things and digital data flows continue to expand exponentially, the
cybersecurity risks from connections through digital networks are rising in tandem.
Increasingly, the ability to cut potentially hostile actors out of the supply chain is
viewed as the most effective means of mitigating the risk of espionage. For example,
the location in mainland China of much of the global semiconductor supply chain’s
assembly, testing, and packaging (ATP) capacity raises the prospect of so-called
“hardware hacks” by Chinese authorities, such as the covert insertion of additional
components for intelligence collection or sabotage.[3] Onshoring the semiconductor
supply chain, or at least “friendshoring” it to countries perceived as friendly, is
increasingly regarded as a precondition for national cybersecurity.
Current policies. Semiconductors were one of four technologies targeted by the Biden
administration’s 100 Day supply chain reviews, a report for which was released in June
2021 (White House, 2021). The report’s conclusions on semiconductors were that U.S.
industry has many strengths along the supply chain, but also critical gaps in
fabrication, manufacturing equipment (lithography), materials, and ATP capacity. The
recommendations included financial support for onshoring initiatives and for the U.S.
semiconductor ecosystem generally, engaging allies and partners to harmonize
semiconductor-related policies and invest in the U.S., and using export controls and
foreign investment reviews to protect U.S. technological advantage and address
national security concerns.
43
The Biden administration has emphasized that the U.S. is in an international
“competition to win the 21st century” through technological leadership. Furthermore,
the debate over industrial policy for semiconductors is taking place in the context of a
growing push for onshoring critical technology supply chains in general. For example, in
2021, Sen. Josh Hawley introduced draft legislation — the “Make in America to Sell in
America Act” — that would impose local content requirements in sectors deemed by
the federal government as critical to national security and protection of the U.S.
strategic industrial base.[4]
The U.S. Department of Defense has long run a “trusted foundry” program for procuring
microelectronics from security-vetted suppliers, but this approach has come under
strain as U.S.-based companies have fallen behind in leading semiconductor
fabrication. The Defense Advanced Research Projects Agency (DARPA) is funding
projects and engaged in public-private partnerships aimed at developing U.S. industry
in fabrication and advanced packaging.[5] DARPA also funds R&D for compound
semiconductor materials — notably silicon carbide and gallium nitride— that have
particular applications and may represent new frontiers for technological development.
U.S. policy measures directed at strategic competition with China in this sector are
commonly typified as “run faster” or “trip the opponent.” On the “run faster” side, in
August 2022 the U.S. passed the CHIPS and Science Act, which provides subsidies for
semiconductor R&D and construction of fabrication plants inside the U.S. These are tied
to “trip the opponent” measures in the form of requirements for subsidy recipients not
to engage in equivalent activities in China, with an exception made for capacity related
to “legacy semiconductors.”[6] This refers to older-generation processes that are
limited to producing chips with lower transistor density, with the CHIPS Act reserving
interpretation of the exact meaning of “legacy” to U.S. authorities.
In September and October 2022, the Biden administration doubled down on a “trip the
opponent” approach with a major expansion of export controls targeting China’s
semiconductor industry and sectors that depend on AI and high-performance
computing. The promulgation notice for the October measures justified them in terms
of restricting Chinese “military modernization, including the development of weapons
of mass destruction (WMD), and human rights abuses.”[7]
In September, National Security Advisor Jake Sullivan framed these steps as a
transition in U.S. strategic technology policy, from staying a couple of generations
ahead of China to maintaining “as large of a lead as possible,” “given the foundational
nature of certain technologies, such as advanced logic and memory chips.”[8] In early
October, U.S. Trade Representative Katharine Tai reinforced this message by
describing China’s industrial policies as a threat to the survival of free societies
44
including the U.S., requiring Washington to prioritize supply chain security above free
trade.[9] As of mid-October 2022, the exact scope and application of the expanded
controls targeting China’s semiconductor sector remained unclear and subject to public
consultation and adjustment by the U.S. Commerce Department, but the key points can
be summarized as follows.
Most of the new rules apply to all entities in China, using a new regulatory category of
“regional stability.” The controls are therefore specific to China and justified by
characterizing China as a threat to regional stability, rather than by the nature of the
controlled items themselves. The new controls target semiconductor manufacturing
equipment (SME) and certain advanced computing ICs and memory chips, focused on AI
and supercomputing applications. Exporting these items and services to an entity
operating in China now requires a license from the U.S. Commerce Department,
application for which is generally subject to a presumption of denial.
Furthermore, non-Chinese entities are restricted from supplying Chinese customers
with the controlled items where their production involves U.S.-origin technology, which
is extensively present throughout the global semiconductor supply chain. Additionally,
one new rule amends the criteria by which U.S. regulators can add an entity to the
“entity list,” which subjects the entity concerned to extended controls and licensing
requirements. These criteria now include “a sustained lack of cooperation by the host
government… that effectively prevents” U.S. authorities from determining compliance
with export controls, in particular end-use checks.[10]
“U.S. persons” are now restricted from working in China in activities related to these
controlled items unless granted a license by the U.S. Commerce Department. This will
affect a significant number of individuals in China’s semiconductor industry who hold
U.S. citizenship and are heavily represented in the startup firms trying to plug gaps in
China’s domestic capabilities. This rule seems to encompass potential offshoring
activity by U.S. entities, to deter them from transferring technology and know-how to
non-U.S. entities that then sell to Chinese customers.
Although nominally confined to advanced semiconductors and related items, these
rules amount to a technological containment strategy, reflecting an expanded concept
of U.S. national security and China’s identification as the primary threat in this context.
In his September speech, Jake Sullivan rejected the distinction between domestic
issues and national security “when facing a competitor that is determined to overtake
U.S. technological leadership” and competing “to lead in the industries of the
future.”[11] While the new rules are justified in terms of military uses and human rights
abuses, the foundational and dual-use nature of advanced ICs and high-performance
computing means that these controls will hamper development of China’s civilian
45
economy on a broad front, particularly in emergent sectors like self-driving vehicles
and intelligent manufacturing that Beijing has identified as priority national
development goals. Furthermore, in practice many of the newly controlled technologies
are not precisely related to the advanced capabilities that are nominally being targeted
but have wider application. This means that supplies to Chinese customers with respect
to older-generation capabilities may be seen as risking violation of the new rules, with
deleterious effects for Chinese industry that are much wider than the framing of the
new controls would suggest.
The new controls took effect immediately and rapidly provoked industry responses.
Within days of the October rules’ promulgation, one of the leading SME suppliers — a
U.S. firm whose largest market is China — reportedly instructed staff that it would
immediately cease offering certain supplies and services to China-based customers,
including third-country companies operating in China like South Korea’s SK Hynix.[12]
However, the presumption of denial for license applications under these controls does
not apply for companies headquartered in specified countries. This reflects
consideration for South Korea and other U.S.-allied governments concerned about
impacts on their economies and technological champions from being forced to
terminate business in and with China in semiconductors. Several foreign industry
leaders including SK Hynix and TSMC have already received one-year licenses under
the new rules to continue operations in China.
A major obstacle to U.S. reshoring efforts, as acknowledged in the Biden
administration’s supply chain review, is that the U.S. is not a cost-effective location for
many activities along the semiconductor supply chain. This has also been highlighted
by TSMC’s former and current leaders in statements about the company’s plans for
operations in the U.S.[13] The sums being debated in Congress are too small to achieve
major shifts in the global supply chain over the short term, when compared against the
industry’s numbers. For example, TSMC is spending $44 billion on capital expenditure
in 2022 alone, while Samsung’s semiconductor foundry division is projected to be
generating over $50 billion in annual revenue by the late 2020s, with some 70% of
revenue currently being reinvested in production capacity.[14]
This makes it unlikely that any U.S. player will be competitive at scale with either TSMC
or Samsung for the foreseeable future, although Intel is attempting to reestablish itself
in the cutting-edge fabrication market. The U.S. government has made efforts to
persuade TSMC and Samsung to locate operations in the U.S., and both have responded
with significant capital investments. But both companies also appear likely to continue
building cutting-edge plants in their home jurisdictions, perpetuating the risks entailed
from the viewpoint of U.S. policymakers.
46
For China, the situation is far more challenging, even before the effects of the October
2022 export controls are factored in. Although Chinese industry has made significant
progress in many steps of the semiconductor supply chain and gained notable market
share in a few segments, Chinese firms are not industry leaders in any step and remain
incapable of producing cutting-edge products in critical niches. It is these gaps that are
being exploited by U.S. export controls that effectively impose secondary sanctions on
third parties for doing business with Chinese companies, which depend on foreign
(including Taiwanese) companies for critical processes such as fabricating cuttingedge chips.
Chinese authorities have been trying for decades to promote development of the
domestic semiconductor industry. In 2014, the national government created a top-level
bureaucratic steering committee (leading small group) and a state-linked investment
fund (the so-called “Big Fund”) to drive development of China’s semiconductor sector.
This represented an effort to harness the private sector’s energies toward strategic
priorities set by state authorities, in line with the “top-level design” approach to
policymaking under Xi Jinping’s leadership.[15]
This approach was premised on China’s participation in the transnational
semiconductor supply chain. Chinese firms were able to become competitive in some
capabilities while relying in other areas on foreign vendors, most notably TSMC and
other foundries for leading-edge fabrication. Chinese firms were also able to build
market share and technical know-how through mergers and acquisitions of foreign
entities, especially in ATP, where mainland China-based companies became the
second-largest bloc by market share, after Taiwanese firms.
Over the past half-decade, the Chinese government has responded to growing
pressures from the U.S. to promote supply chain “decoupling” from China by doubling
down on import substitution efforts. Over 2020-2021, Chinese national agencies
introduced three sets of measures providing targeted support for the semiconductor
sector that include tax relief, direct financing and subsidies, regulatory guidance, and
skills development. Semiconductors were one of seven frontier technologies prioritized
by the 14th Five-Year Plan released in March 2021. This signal of strategic importance
led to a proliferation of semiconductor industry development plans by provincial and
municipal governments across China.
Policy emphasis seems to be moving to advanced packaging techniques and compound
semiconductor materials, being fields in which Chinese industry has some prospect of
leapfrogging the gap with foreign industry leaders. Compound materials were the only
reference to semiconductors in China’s 14th Five-Year Plan for National Informatization
47
released in January 2022. They also feature prominently in policies being released by
subnational governments, notably in the Shanghai government’s ICT sector
development plans, which include the ambition for a Silicon Carbide Valley” industrial
cluster.
All these new policies emphasize developing a complete semiconductor ecosystem,
relying on synergies between companies at different steps of the supply chain. Political
rhetoric increasingly exhorts a “whole-of-nation” approach to reducing China’s
weaknesses in critical technologies, drawing on a mobilizational tradition that goes
back to the strategic defense projects of the Maoist era. Chinese authorities may be
developing a successor to the ’02 Special Project, a decade-long program launched in
2009 that targeted priority semiconductor-related technologies for development by
combined efforts from Chinese companies and research institutions. However, a new
round of such specific cross-industry R&D goals has yet to be publicized.
Overall, China remains heavily reliant on foreign inputs in the semiconductor sector,
reflected in the well-known statistic of the nation’s semiconductor imports now
exceeding the value of its oil imports. But this statistic also reflects the concentration
of global electronics manufacturing in China, which has so far provided strong
incentives for foreign firms to remain engaged with Chinese markets and leverage for
Chinese authorities: Xi Jinping has emphasized the need to ‘pull tight’ global supply
chains to China.
Since the new U.S. export controls of late 2022 are confined to certain categories of
advanced semiconductors and items required to produce them, large sections of the
Chinese semiconductor industry should be able to remain integrated with global supply
chains and markets. But unless Chinese industry can quickly substitute at scale for the
technologies controlled by the new U.S. rules, either by developing them domestically
or by procuring them from third countries, China will lose its potential first-mover
advantage in many of the emerging civilian and military applications that its
semiconductor policies were intended to support.
Proposals for finding firmer ground. The policy approach represented by the CHIPS
and Science Act, and especially by the expansion of U.S. export controls in September
and October 2022, has for the time being removed any prospect of finding common
ground between the U.S. and China on strategic policy for semiconductors. The Biden
administration now appears committed to a goal of containing China’s development in
this sector and in strategic technologies built upon it, without clear distinction in
practice between civilian and military applications. This is unlikely to change with a
future return to a Republican administration. For its part, China’s government has
characterized the expanded controls as designed to maintain U.S. “technological
hegemony” and “to hobble and suppress the development of emerging markets and
developing countries.”[16]
48
An important factor will be the willingness of third-party countries to cooperate and
coordinate with U.S. measures. To date, the U.S. has had limited success in bringing
allied governments and their leading companies on board with U.S. policy initiatives for
the semiconductor sector. The U.S. Commerce Department’s 2021 request for
information from actors throughout the global semiconductor supply chain — which
was accompanied by implied threats to compel cooperation from foreign companies if
necessary — was poorly received in Taiwan and South Korea, where it was seen as
overbearing and creating risks of proprietary information leaking to U.S. competitors.
Discussions over a putative U.S.-led “Chip 4 Alliance” with Japan, South Korea, and
Taiwan also did not appear to have delivered any significant outcomes as of midOctober 2022.
European ambitions for “technological sovereignty,” which are clearly expressed in the
EU Chips Act proposal of early 2022, are likely to impose constraints on coordination
with Washington. At a press briefing in October on the new semiconductor export
controls, U.S. officials conceded that they had not secured any promises that allied
nations would implement similar measures and that discussions with those nations
were ongoing.[17] Conversely however, the importance of the U.S. economy and the
amount of U.S.-origin technology present throughout the global semiconductor industry
means that in many cases, attempting to circumvent U.S. export controls to continue
business with Chinese customers would be highly risky.
Furthermore, the risk of doing business with Chinese entities has been amplified
beyond presently controlled items by the amendment to criteria for addition to the U.S.
entity list. As one commentator summarized this change, in sectors targeted by U.S.
controls, “Any company in China can be cut off from worldwide supply chains … through
the justification that China does not cooperate with the U.S. (regulators).”[18] The only
way to avoid this is unreserved submission to U.S. government demands for information
disclosure. It is unlikely that any Chinese firm will take such an approach, which would
run counter to Chinese domestic political imperatives and data security regulation, and
on account of the latter would likely violate specific Chinese laws.[19]
The sweeping new export controls targeting semiconductors and the policy framing
provided by the national security advisor suggest that the U.S. government has reached
a consensus that at the advanced end of these technologies, “hard decoupling” and
restricting Chinese capabilities must be the priority, even if this involves significant
harm to U.S. industry. Furthermore, Washington is pursuing this policy unilaterally in
the acknowledged absence of cooperation by allied countries, in the hope that they can
be persuaded to adopt similar measures and so effectively align against China in their
future economic and technological development. Given the foundational nature of
semiconductors, the new U.S. controls create the conditions for a true bifurcation in
49
global technology ecosystems. It remains to be seen whether third parties will find
these new conditions sufficient reason to reverse the progress of economic integration
with China that for many countries has been a consistent trend for the past quarter
century.[20]
50
Notes:
[1] Jan-Peter Kleinhans. 2020. ‘The Global Semiconductor Value Chain: A Technology Primer for Policy
Makers.’ Stiftung Neue Verantwortung. https://www.stiftung-nv.de/de/publikation/globalsemiconductor-value-chain-technology-primer-policy-makers
[2] National Security Commission on Artificial Intelligence. 2021. Final Report, National Security
Commission on Artificial Intelligence, https://www.nscai.gov/wp-content/uploads/2021/03/Full-ReportDigital-1.pdf
[3] John Lee and Jan-Peter Kleinhans. 2021a. ‘Mapping China’s semiconductor ecosystem in global
context: Strategic dimensions and conclusions.’ Mercator Institute for China Studies and Stiftung Neue
Verantwortung. https://merics.org/en/report/mapping-chinas-semiconductor-ecosystem-global-contextstrategic-dimensions-and-conclusions
[4] Josh Hawley. 2021. ‘The Only Way to Solve our Supply Chain Crisis is to Rethink Trade’. New York
Times. https://www.nytimes.com/2021/10/29/opinion/hawley-supply-chain-trade-policy.html
[5] Defense Advanced Research Projects Agency. 2021. ‘DARPA Joins Public-Private Partnership to
Address Challenges Facing Microelectronics Advancement’. https://www.darpa.mil/news-events/2021-
12-22
[6] Paul Triolo. 2022. 'U.S. finally passes semiconductor subsidy bill, but it’s not going to help competition
with China'. The China Project. https://thechinaproject.com/2022/08/18/u-s-finally-passessemiconductor-subsidy-bill-but-its-not-going-to-help-competition-with-china/
[7] See 2022-21658.pdf (federalregister.gov). Bureau of Industry and Security 15 CFR Parts 734, 736,
740, 742, 744, 762, 772, and 774, Implementation of Additional Export Controls: Certain Advanced
Computing and Semiconductor Manufacturing Items; Supercomputer and Semiconductor End Use; Entity
List Modification.
[8] See Remarks by National Security Advisor Jake Sullivan at the Special Competitive Studies Project
Global Emerging Technologies Summit. The White House. https://www.whitehouse.gov/briefingroom/speeches-remarks/2022/09/16/remarks-by-national-security-advisor-jake-sullivan-at-the-specialcompetitive-studies-project-global-emerging-technologies-summit/
[9] See Remarks by Ambassador Katherine Tai at the Roosevelt Institute's Progressive Industrial Policy
Conference. Office of the United States Trade Representative. https://ustr.gov/about-us/policyoffices/press-office/speeches-and-remarks/2022/october/remarks-ambassador-katherine-tai-rooseveltinstitutes-progressive-industrial-policy-conference
[10] See Department of Commerce Bureau of Industry and Security 15 CFR Part 744 'Revisions to the
Unverified List; Clarifications to Activities and Criteria that May Lead to Additions to the Entity List'.
https://public-inspection.federalregister.gov/2022-21714.pdf
[11] See Remarks by National Security Advisor Jake Sullivan at the Special Competitive Studies Project
Global Emerging Technologies Summit. The White House. https://www.whitehouse.gov/briefingroom/speeches-remarks/2022/09/16/remarks-by-national-security-advisor-jake-sullivan-at-the-specialcompetitive-studies-project-global-emerging-technologies-summit/
51
[12] Josh Horwitz. 2022. 'Exclusive: KLA to stop sales and service to China to comply with U.S. export
curbs'. Reuters. https://www.reuters.com/world/china/exclusive-kla-stop-sales-service-china-complywith-us-export-curbs-source-2022-10-11/
13] Charlie Campbell. 2021. ‘Inside the Taiwan Firm That Makes the World’s Tech Run.’ TIME.
https://time.com/6102879/semiconductor-chip-shortage-tsmc/
[14] Bogdan Solca. 2022. ‘Samsung plans to overtake TSMC by 2030.’ NotebookCheck.
https://www.notebookcheck.net/Samsung-plans-to-overtake-TSMC-by-2030.593861.0.html
[15] John Lee and Jan-Peter Kleinhans. 2022. ‘Europe’s dependence on Chinese semiconductor
manufacturing’. In Digital Power China Research Consortium. ‘China’s Digital Power: Assessing the
Implications for the EU’.
https://timruhlig.eu/ctf/assets/x93kiko5rt7l/4uiZoNQtRkni5KfuNDrBbx/fd52e3320cfe21e6b304ad31d8
1279d8/DPC-full_report-FINAL.pdf
[16] Mark Magnier. 2022. 'Tech war: Washington takes new steps to frustrate China, advance US chipmaking | South China Morning Post'. South China Morning Post.
https://www.scmp.com/news/china/diplomacy/article/3195254/tech-war-washington-takes-new-stepsfrustrate-china-advance-us
[17] Stephen Nellis et al. 2022. 'China and USA Are Officially At Economic War – Technology Restriction
Overview – SemiAnalysis'. Reuters. https://www.reuters.com/technology/us-aims-hobble-chinas-chipindustry-with-sweeping-new-export-rules-2022-10-07/
[18] Dylan Patel. 2022. 'China and USA Are Officially At Economic War – Technology Restriction
Overview'. Semianalysis. https://www.semianalysis.com/p/china-and-usa-are-officially-at-economic
[19] John Lee. 2022. 'Cyberspace Governance in China: Evolution, Features and Future Trends'. Institut
français des relations internationales. https://www.ifri.org/en/publications/notes-de-lifri/asievisions/cyberspace-governance-china-evolution-features-and-future
[20] John Lee. 2021. ‘The Internet of Things: China’s Rise and Australia’s Choices’. Lowy Institute.
https://www.lowyinstitute.org/the-interpreter/china-australia-internet-things
52
VIII. Opportunities for Dialogue
Challenges and opportunities. As the sections above have outlined, economic, political,
and social factors shape the space for U.S.-China dialogue on the future of technology.
Technology is a suite of tools embedded in economic activity, law enforcement, defense
modernization, and social space. The ways a government regulates and promotes
technological development is unique to its political and market system — and between
the U.S. and China, these systems’ differences contribute to mutual strategic mistrust.
Each side has different preferences for the role of the state in the market. The U.S. has
moved over several decades to deregulate most industries on the premise that market
decisions are more efficient and profitable than state intervention. China has cycled
through periods of loosening and tightening state control of the market to balance the
benefits of market freedom against the political and economic risks of wealth and
power bubbles in an authoritarian one-party system. When the U.S. argues that China is
creating an unfair playing field for normal economic competition, it is often due to
policies that China sees as necessary for its long-term political economy. A frank and
authoritative dialogue on each side’s drivers of tech policy decision-making would help
contextualize policy developments and test assumptions about each side’s ultimate
goals; however, system differences also distort the opportunities for a productive
dialogue.
Each side has different standard operating procedures for bilateral dialogue and
diplomacy. The U.S. changes personnel and priorities with incoming administrations and
typically allows working-level officials wide latitude to negotiate within the political
mandate (i.e., to explore what is possible). Chinese officials tend to stay long-term in
one issue area but often have very little room to suggest policy changes without
explicit top-down direction. These differences have led to negotiation fatigue between
the two sides, particularly in the absence of a wide-ranging diplomatic process that
filled these gaps.
Both sides have perceptions of the risk environment that increase bilateral friction. Put
simply, Chinese technology is seen in the U.S. as an extension of party-state assets and
technological cooperation as aiding a competitor or rival. As China trends toward
tighter party-state control of the private sector, the U.S. evaluates new Chinese
regulations on their potential to create strategic vulnerabilities and new U.S.
regulations on their potential to defend against these vulnerabilities. Meanwhile, China
views the U.S. emphasis on personal freedom and electoral democracy as a strategic
risk in adopting U.S. technology and social platforms.
53
A roadmap to a productive bilateral dialogue between the U.S. and China on emerging
technology must manage the following challenges:
1. Keeping expectations realistic. Technology’s dual-use applications and
pervasiveness in everyday life has exponentially widened the national security
community’s perception of risk in U.S.-China cooperation. The amount of dialogue
necessary to let technology discussions drive better bilateral understanding and
cooperation is unlikely to materialize. Instead, fewer dialogues should prioritize highlevel management of the key priority issues: the role of AI in warfare, IoT
standardization, ethics and norms in data collection, and fair use of cyber tools.
The output of these dialogues might be limited to reducing misperception by
investigating the other side’s evolving position on these topics. However, regularized
dialogue may identify sufficient common interests — for example, requiring human
confirmation of AI decision-making in warfare — that can lead to formal negotiations
and agreements. IoT standardization discussions can address concerns about backdoor
capabilities built into objects by SOEs and national champions.
Keeping expectations to issue management may forestall use of the dialogues
themselves as leverage, as when the U.S.-China Cyber Working Group (CWG) was cut
off by Beijing in the aftermath of the U.S. indictment of five Chinese military officers on
cybertheft charges. Though a high-level dialogue was reinstated through a 2015 XiObama summit joint statement, this dialogue was only held once more after being
renamed in the Trump administration. Reaffirming the five principles reached in the
Obama-Xi agreement again in the Biden administration would be a stabilizing factor in
bilateral relations and proof that U.S. agreements can be sustained throughout swings
in domestic politics.
U.S.-China military-to-military dialogues provide a useful analogy and point of
optimism. After China cut off military-to-military dialogues many times over U.S. arms
sales to Taiwan, these dialogues had become useful and necessary mechanisms that
survived the breakdown of other diplomatic dialogues after the 2017 U.S. National
Security Strategy emphasized great power competition as a guiding principle of U.S.
strategy toward China. Issue management, therefore, seems to be an area where
dialogue is both possible and stabilizing.
A focus on issue management does not preclude agreements on underlying principles
in new areas. While more generalized and likely less enforceable than desired, coming
to bilateral agreement on basic principles can be seen as a component of issue
management and serve to deepen understanding on how the two sides see the major
54
issues from their own political and historical context. The point would not be to change
the other side’s view, but to see at what level the two sides’ interests overlap, no matter
how general.
2. Getting the right people together. Because the use of technology cuts across social,
economic, and military paradigms, many different government agencies participate in
the policymaking process. For example, the charter of the U.S. National Science and
Technology Council Select Committee on Artificial Intelligence mandates membership
from 10 different government agencies or sub-agencies.[1] In China, tech policy often
overlaps among a similar number of agencies, including some that have no direct
corollary in the U.S. system. To compare on AI rules-making, China has three bodies—
the Cyberspace Administration of China, China Academy of Information and
Communications Technology, Ministry of Science and Technology — contributing
frameworks and rules to an overall governance regime.[2]
A first step in initiating a productive issue-management dialogue between the U.S. and
China is to identify the correct interlocutor on the technology issue in question. With
whom should the Ministry of Science and Technology personnel communicate in the
U.S. system, which has no equivalent Cabinet department? Previous large-scale
engagement efforts, such as the U.S.-China Strategic and Economic Dialogue,
successfully navigated these various divisions of labor and responsibility by
establishing robust working-level relationships. A similar process would need to
develop to put the right people together on technology policy issues.
U.S. officials may also need to seek input from or incorporate feedback from privatesector actors into the dialogue. In the current political climate, joint research seems out
of the question. However, the moving target of technological innovation and the
specific issues that could be remedied through dialogue and consultation — supplychain bottlenecks, cybersecurity concerns, and so on — requires the private sector to
periodically weigh in. China closes this knowledge gap through several formal
mechanisms that reduce the space between state and industry; the U.S. needs to make
sure that its structure of government-industry knowledge-sharing is likewise robust.
3. Setting the agenda. Both countries are seeking a competitive edge over the other
and will not want to share specifics on their progress, particularly as new technologies
and tools are developed. A workable agenda for U.S.-China tech dialogue would cover
the broadest and the narrowest issues and leave the space between off-limits; for
example, the two sides can discuss rules for fair use of cyber tools in peacetime
(broad), or resolve specific customs issues (narrow), but are unlikely to seek renewed
research collaboration on sensitive technologies.
55
One component of bilateral dialogue and exchange on technology issues must be each
side’s interpretation of rules of fair use of cyber tools. With the dual-use capabilities
described in the sections above, the U.S. and China are likely to have some degree of
mutual vulnerability to cyberattacks, potentially ones that target big data sets, use AI
capabilities and/or occur over 5G networks. A sound deterrence strategy requires each
side to know exactly what is escalatory or unacceptable to the other side. By defining
these parameters, both sides can then be clear on what practices are perceived as
outside the bounds of fair use and therefore risk retaliation.
Relatedly, the two sides can continue to discuss their different approaches to data
storage and privacy, though neither side should expect to change the fundamental
position of the other on these issues. Through discussion, both sides should recognize
that their differences lie in their respective justice systems and not in the tools
themselves.
Supply-chain issues, particularly those related to semiconductors, will not be a
productive area for bilateral dialogue, but discussions on macroeconomic stability are
of common interest and could touch on the disruptions associated with bifurcating the
technology supply chain as well as any accusations of weaponized interdependence
regarding the necessary minerals and elements in technological production.
4. Committing to a regular schedule. Regularization of a dialogue can mitigate certain
challenges. First, the rapid pace of innovation requires constant agenda adjustments on
which one or both sides may be reluctant to agree. Regular meetings can help each side
understand one another’s views and trajectory, and work through potential conflict
regarding misperceptions of technology policies.
Second, a regular schedule creates an action-forcing mechanism for both systems to
conduct internal reviews and formalize policy. Since each side will be required to
discuss technology-related issues with one another, internal plans must be consistent
with dialogue talking points, requiring each state to understand the positions of its
different stakeholders and ensure policy guidance is in place.
Third, a regular schedule builds relationships. This is particularly important because
U.S. officials can change from administration to administration. Regular meetings can
allow both sides to assess the personalities and interests of individual participants and
to strengthen ties among them. Because of the differences in bureaucratic
responsibilities outlined above, these relationships may also prove useful on issues not
covered in the technology dialogues.
56
Finally, a regular schedule allows a long-term approach to bilateral discussion on
technology issues. Many of these issues cannot be resolved or shelved. Knowing that
another meeting is already on the calendar reduces pressure to prematurely declare
the failure of dialogue and diplomacy; as a result, gains can snowball over time.
5. Managing the interplay between bilateral discussions and international institutions
and agreements. The U.S. and China each have considerable weight in the international
system. Where bilateral agreements are in place — such as the Obama-Xi principles for
cybersecurity — the two countries could push for expanded adoption of such principles
in larger multilateral formats. In other words, the bilateral agenda should be the
starting point for establishing global rules and norms. Bringing the rest of the world on
board with relatively benign principles, such as timely responses to official inquiries on
cyber activity or that no government actors should participate in the theft of
intellectual property, would reinforce major power accountability to these tenets.
A discussion on rules can not only stabilize the bilateral relationship by putting
guardrails on the use of technology, but also pave the road for forward-leaning
discussions with other major powers and global actors. Both the U.S. and China are
sufficiently powerful to ignore rules imposed by one side on the other; the only
enforceable rules-based order on technology tools will be reached via dialogue and
negotiation.
Where agreements — even on basic principles — are often impossible, multilateral
mechanisms can expand and supplement the conversation. Outside of U.S.-China
strategic competition, rules and regulations in the EU areas are driving standards and
norms on data privacy. And, as suggested above, a global governance framework or
institution that can address cyberconflict should consider the U.S. and Chinese
positions on technology rules and norms. However, any such framework must consider
the very high start-up costs —including time, political buy-in, and money — associated
with creating a new institution as well as the fact that current institutions face
significant obstacles given heightened major power competition. Any institution will
find it incredibly difficult to function in the absence of bilateral agreements or
multilateral agreements that include both Washington and Beijing. Already, the
dominant trend is toward using international institutions as an arena of major power
competition rather than as a venue for consolidating common interests. Perhaps this
may have to be an area for tentative, early exploration.
57
Notes:
[1] Trump White House. 2018. Charter of the U.S. National Science and Technology Council Select
Committee on Artificial Intelligence, 2018) Trump White House Archives, Jan 5,
https://trumpwhitehouse.archives.gov/wp-content/uploads/2021/01/Charter-Select-Committee-on-AIJan-2021-posted.pdf
[2] Matt Sheehan, 2022, China’s New AI Governance Initiatives Shouldn’t Be Ignored. Carnegie
Endowment Commentary, January 4, https://carnegieendowment.org/2022/01/04/china-s-new-aigovernance-initiatives-shouldn-t-be-ignored-pub-86127
58
IV. Conclusion
Greater transparency, discussion, and cooperation are critical components of reducing
the risk of technology conflict. Even though the U.S. and China must treat technologies
as areas of high national security interest, both countries should recognize that the new
technology landscape does not need to derail their common concerns. Technology is a
suite of tools that compound underlying dynamics, integrated into economic
competition, defense planning, and political values while having potentially profound
consequences for society. Some of the effects of new technology — both direct and
indirect — are not yet fully fleshed out, much less understood. However, these tools are
not separate from existing U.S.-China dynamics but evolve in the context of political,
economic, and security policy decisions.
Restoring stability in U.S.-China relations will require long-term and enduring
management of emerging technology concerns. The U.S. and China should approach
each other on these issues with cool heads and well-defined bottom lines. The purpose
of technology dialogue is to explore where political and economic interests on
technology trade and cooperation overlap, and where such interests are intractable.
Technology dialogue will be most productive if it bridges gaps between the U.S. and
Chinese bureaucracies and if the officials involved avoid invective while arguing for
their own policies. It will be least productive if either side approaches with the goal of
airing grievances about the other side’s political system and economic interests.
The U.S.-China relationship was strong for many years until recent events prompted a
reframing of the relationship from cooperative to competitive and confrontational. The
cooperative aspects of the relationship can be rebuilt if both sides are willing to enter
into good-faith dialogue, even if a political climate that allows for such exchanges to
gain traction takes time to develop. The consequences of avoiding such dialogue could
include prolonged global economic inflation due to decoupling and threat perceptions
that could escalate regional and global instability. In the current international
environment, there is little appetite for either. The U.S. and China should avoid costly
mistakes that could result from refusing to talk about technology interests. Regardless
of the trajectory of U.S.-China relations, holding out the possibilities for interaction and
even potential cooperation remains an important aspect of the relationship.
59
About the Publisher
This report is published by the Carter Center's China Focus. Its editors are Yawei Liu,
senior advisor on China at The Carter Center, and Michael Cerny, program associate for
the Center's peace programs.
President Carter's decision to normalize diplomatic relations between the United
States and the People's Republic of China in 1979 changed both countries and the
world. Facilitating expanded bilateral trade, investment, and people-to-people
exchange between the two countries has allowed East Asia to enjoy relative peace and
prosperity for decades.
However, the U.S.-China relationship is now under immense strain. Since 2009, the
Chinese Communist Party has strayed from the path of “reform and opening” that
encouraged slow and steady progress toward political and economic liberalization.
Washington began to criticize China’s attempts to revise the international system, and
Beijing responded by accusing the U.S. of containing China’s rise. As President Carter
wrote in February 2021, “Government officials in both countries have adopted rhetoric
and policies that reflect the hostility that Vice Premier Deng and I sought to calm in
1978.”
The Carter Center remains committed to preserving the legacy of President Carter and
Deng Xiaoping’s historic decision while adapting to the demands of the 21st century.
This requires navigating a bilateral relationship fraught with global crises, ideological
divergence, human rights crises, nationalist tension, and the looming threat of conflict
in the Taiwan Strait. Through its research, workshops, and online engagement
initiatives, the China Focus fosters greater dialogue, exchange, and critical reflection
on the past, present, and future of U.S.-China relations.
Contact: In Atlanta, Maria Cartaya, maria.cartaya@cartercenter.org
60
