DON'T LOSE ACCESS:
Your IP access to ForeignPolicy.com will expire on June 15.
To ensure uninterrupted reading, please contact Rachel Mines, sales director, at firstname.lastname@example.org.
China’s AI Giants Can’t Say No to the Party
Open debate about the ethics of tech is a strength, not a weakness, of the U.S. system.
As China seeks a deeper fusion between defense and commercial technological development, the bridges that the U.S. Defense Department has been seeking to build to Silicon Valley seem precarious. Google faced intense backlash against its work with the Pentagon’s Project Maven, ultimately deciding not to renew the contract—even though this engagement appears to be consistent with the ethical principles that Google has since released. When ethics collide with strategic competition in artificial intelligence, the United States can seem to be at a disadvantage.
In contrast to the resistance from Silicon Valley in the United States, China is stepping up a national strategy of “military-civil fusion” (also translated as “civil-military integration”) that concentrates on creating and leveraging ways to cooperate on these dual-use technologies, enlisting technology companies and universities, including tech giant Baidu and Tsinghua University, to promote their military applications. It might seem as though China has the edge here, given Beijing’s power to command the tech sector. But the Chinese system also has critical weaknesses—and the U.S. one has unexpected strengths.
The recent controversies reflect Google’s international nature. While its headquarters are in the United States, Google employees come from a growing number of countries. Google has established AI research centers in Paris, New York, and Tokyo, as well as in Beijing last winter, with one in Accra, Ghana, opening soon. That international character is a unique strength and competitive advantage—particularly in the global quest for AI talent—but inherently complicates arguments that Google should support the U.S. military and advance U.S. national security interests.
The campaign against the company’s involvement in the development of warfare technology reflected globalized engagement and mobilization in response to intense concerns and controversy over the weaponization of AI.This activism at Google is occurring at a time when tech workers in the United States are actively organizing to resist and protest, spurred by intense concerns over the policies of the Trump administration. In particular, U.S. staff are demanding that Microsoft, Palantir, and other tech companies cancel their contracts with U.S. Immigration and Customs Enforcement to prevent the abuse of their technologies to enable cruel policies and human rights violations.
Google’s secret development of a search engine that it plans to launch in China—which will comply with censorship and will “blacklist sensitive queries,” such as searches about issues of human rights and democracy—may also confront strong resistance from the company’s workforce.
These open debates—and the freedom to challenge authority—are integral to and inextricable from the creativity and risk-taking that are vital to innovation in U.S. technology ecosystems such as Silicon Valley.
It is difficult to imagine a similar campaign could have occurred, let alone succeeded, in China. Even though leading Chinese technology companies are starting to become international in their presence, workforce, and activities, the Chinese Communist Party (CCP) has reasserted its authority over them. Chinese President Xi Jinping has declared that the party “leads everything,” a troubling re-emergence of a Mao-era sentiment. The CCP has expanded its presence within major tech companies; a majority of them even have party secretaries who represent the CCP’s interests and authority, including the “big three” of Chinese tech: Baidu, Alibaba, and Tencent.
The CCP has also sought to co-opt the leaders of tech companies into the structure of the party itself, including through their inclusion in United Front Work organizations such as the Chinese People’s Political Consultative Conference.
This spring, in an uncharacteristic (and later censored) moment of candor, the CEO of Sogou, a tech company known for its search engine and development of speech recognition technologies, declared, “We’re entering an era in which we’ll be fused together. It might be that there will be a request to establish a Party committee within your company, or that you should let state investors take a stake … as a form of mixed ownership. If you think clearly about this, you really can resonate together with the state. You can receive massive support. But if it’s your nature to go your own way, to think that your interests differ from what the state is advocating, then you’ll probably find that things are painful, more painful than in the past.”
China’s “national champions,” as well as start-ups and emerging enterprises, in AI are becoming ever more closely linked to party-state priorities, including development of surveillance capabilities. For instance, iFlytek—known for creating “China’s Siri”—which recently established a partnership with MIT, has also been involved in the development of surveillance capabilities in Xinjiang that leverage its smart voice technologies. There are numerous examples of facial recognition companies, such as Yitu Tech and SenseTime, that directly support policing and public security within a system in which these capabilities are often abused.
At the same time, as the Chinese People’s Liberation Army pursues a range of military applications of AI, leading universities and enterprises are actively supporting these developments. Tsinghua University, often characterized as China’s MIT, has highlighted its commitment to supporting China’s national strategy of military-civil fusion, including establishing the Military-Civil Fusion National Defense Peak Technologies Laboratory and building the High-End Military Intelligence Laboratory with support from the Central Military Commission. Meanwhile, Baidu has partnered with a research institute from the China Electronics Technology Group Corporation, a state-owned defense industry conglomerate, to create the Joint Laboratory for Intelligent Command and Control Technologies, which will focus on using big data, artificial intelligence, and cloud computing to enhance military command information systems.
So far, there have not been any indicators of resistance to the idea that Chinese technology companies should be in service of the party-state. That’s hardly surprising; Xi’s regime has cracked down harshly on dissent, and open policy debates are far more limited than they were even just five years ago.
There are discussions on AI ethics taking place in China, nevertheless. There’s a range of very active conversations, including statements from prominent philosopher Zhao Tingyang, who has argued forcefully against the development of superintelligence capable of “saying no” to humans; Zhou Zhihua, the only scholar from mainland China to sign a letter calling for a boycott of a South Korean laboratory believed to be working on military applications of AI; and Chinese Army defense academic Zhu Qichao, who has discussed the risks of AI in military affairs.
Such engagement with these issues in China should be welcomed. But it’s possible that the Chinese government may attempt to co-opt these conversations to assert its own centrality, in ways that advance party-state priorities. For instance, the New Generation Artificial Intelligence Development Plan calls for China to lead in AI ethics, standards, and global governance.
Moreover, it seems unlikely that the open conversations—in which civil society takes the lead in mobilizing and creating partnerships among a diverse range of researchers and stakeholders—that are occurring within the United States and around the world could happen in Xi’s China. At the core of this activism is a concern with and challenge to the ways in which militaries and governments may use and abuse AI that could infringe upon civil liberties and human rights. In China, any mobilization that is perceived as a potential threat to the party’s own security, whether by lawyers, feminists, or veterans, tends to be rapidly, even violently, suppressed. So too, any concerted attempts to organize against government surveillance seem unlikely to be tolerated.
At a time when China’s ambition to “lead the world” in AI is provoking alarm and enthusiasm, and as great power competition intensifies, there are moments when China’s capacity for state control and direction may seem appealing or even enviable—especially for strategists frustrated by the unwillingness of idealists in Silicon Valley to support the U.S. military directly. Yet there are reasons to doubt the Chinese strategy. So far, in China’s AI revolution, tech companies have been at the forefront, and this “national team” will be integral to the realization of China’s AI ambitions. While the dynamism and relative independence of these companies have been critical to their success, their power is perceived by the party as a threat, which may lead to some painful outcomes.
The CCP’s prioritization of control may result in inherent contradictions. At a time when China is pursuing indigenous innovation, the expansion of the CCP’s presence within tech companies—which are now expected to promote the implementation of “Xi Jinping thought”—may harm creativity and innovation. As companies’ connections to the CCP and Army start to provoke concerns, this entanglement may also impede their activities overseas.
There are also reasons for skepticism about the future trajectory of China’s grandiose AI plans. While this state-driven approach seeks to harness the dynamism of tech companies and concentrates on building up a major market, these policies could result in inefficiencies, or even exacerbate the risks of an AI bubble through excessive investment.
There’s also a certain amount of political bandwagoning going on, as provincial governments seek to play follow-the-leader with Xi’s priorities. The growing number of local governments that have issued their own AI plans are unlikely to all succeed.
The United States must recognize its own enduring advantages, including the dynamism and inclusivity of its innovation ecosystems. The open and intense debates over AI ethics should not be dismissed as a potential disadvantage—but rather recognized as integral to America’s values and vitality as a democracy.
Going forward, the United States should reaffirm its commitment to ensuring that AI technologies are developed and deployed in ways that are consistent with the country’s liberal and democratic values. In the process, it should support an inclusive, multi-stakeholder approach that highlights the voices and concerns of civil society in policy debates on AI safety, standards, and governance. Despite the imperfections of and current challenges to its own democratic institutions, the United States may yet revitalize its own democracy and leadership in AI.