Smart Money on Chinese Advances in AI



William A. Carter and William Crumpler | CSIS

Artificial intelligence (AI) has become a central theme of global strategic competition. AI promises unprecedented gains in economic and military power to countries that can effectively harness its potential to derive insights from the proliferation of global data. Few countries have embraced the vision of an AI-powered future as fervently as China. China’s AI ecosystem is very different from that of the West and remains largely opaque to foreign observers. Unlike the United States, the Chinese government is dedicating significant resources and attention to AI development and creating a supportive policy environment to facilitate innovation and experimentation and proactively manage risk.

Numerous misconceptions and competing narratives around China’s innovation economy have made it difficult for U.S. policymakers to understand the AI ecosystem in China and its links to AI innovation in the United States. This report seeks to improve this understanding by examining China’s progress toward achieving its four strategic goals. The centerpiece of China’s national AI strategy is the New Generation Artificial Intelligence Development Plan (NGAIDP), released by the State Council in 2017. Part of China’s ambitious Made in China 2025 (MIC2025) initiative, NGAIDP outlines a series of strategic initiatives to make China the world leader in AI by 2030. The plan highlights the four strategic goals China seeks to achieve in its pursuit of global AI leadership:

1. Make China the center of fundamental AI research and AI theory development;
2. Lead applied research and development of cutting-edge AI products and services;
3. Build a world-leading domestic AI industry and
dominate global markets for AI; and
4. Leverage AI to increase efficiency in traditional industries and move up the global value chain.

But China’s approach to AI differs from traditional Chinese state planning. China’s high-tech companies are a new breed of Chinese state champions, not having grown up under the wing of the Chinese Communist Party (CCP). In contrast to other industries like banking or telecommunications, companies like the BATs (Baidu, Alibaba, and Tencent) are not former state-owned enterprises (SOEs) and lack the formal links to the CCP that state champions in other industries possess. As Chinese officials and technologists repeatedly emphasize, NGAIDP, MIC2025, Internet+, and other Chinese industrial policies for high-tech industries serve as guidance and signaling mechanisms to the private sector rather than direct government control Chinese leaders recognize that achieving the four goals requires more than just investment in fundamental research. Rather, they utilize an ecosystem approach, combining workforce development, fundamental and applied R&D, deployment of enabling technologies and infrastructure, and incentives for AI adoption in traditional industries. China’s progress in realizing these goals has become a subject of intense interest as the U.S.-China relationship comes to be increasingly defined by technological competition.

China has made significant progress toward developing homegrown AI talent but still struggles to retain the talent it produces. The Chinese government has helped expand AI programs at universities across the country, making China the world’s second-largest home for AI talent, and has utilized the Thousand Talents Program to draw technologists and researchers to Chinese institutions. However, China continues to struggle to retain elite talent, who tend to prefer U.S. and global institutions because they provide the opportunity to work with global leaders in their field without any of the concerns over intellectual independence, air quality, food safety, or educational quality for their families.
Aggressive investment in AI R&D from both the public and private sectors has helped drive significant progress in China’s AI research and industrial bases, but producing unique research and cutting-edge technologies remains a challenge for China. The United States is still responsible for the highest-impact research, but Chinese scholars are increasingly gaining prominence in the field. Focused investment in strategic AI verticals like autonomous vehicles, facial recognition, and natural language processing has helped to boost industry players in China, and the development of digital platforms by internet giants like the BATs is driving investment in new AI tools and applications.

The development of enabling technologies and infrastructure for AI has benefited from China’s ecosystem approach to AI investment. Major national projects to promote investment in key technologies like the internet of things, 5G networks, cloud computing, and robotics are helping China to advance its AI goals by spurring the growth of the requisite physical and digital infrastructure that success in AI will
depend on.

Finally, while China is working to spur AI adoption in traditional industries, the country faces a number of significant challenges. In some areas, such as consumer apps and digital services, Chinese companies have developed world-leading applications, leveraging their massive consumer base and bundled app ecosystems to build enormous databases on Chinese consumers. But AI’s true potential lies in unlocking significant productivity gains and competitiveness through the integration of AI into enterprises and traditional industries. Low rates of digitization among Chinese enterprises, barriers to accessing public data, lagging cloud adoption, uncertain data quality, lack of data diversity, and protectionist data policies make that difficult. Until these issues are rectified, AI developers in China will struggle to create AI applications that produce value for China’s traditional industries.

Overall, China’s progress toward AI leadership remains uneven, and the global AI ecosystem, led by the United States and key allies like the United Kingdom and Canada, continues to attract the top talent, produce the highest-quality research, and be best-positioned to capture the greatest value from AI deployment throughout the global economy. However, China is utilizing every available lever of state power and investing significant attention and resources to develop its AI ecosystem and become a leader in AI innovation. Led by private companies and researchers and supported by significant funding and enabling policies from the Chinese government, China will almost certainly continue to grow its contributions to AI research and application development, expand its AI industries and increase its exports of intelligent products and services, and perhaps even overcome its legacy challenges to increase Chinese competitiveness in traditional industries.

While many policymakers in the United States react with alarm to China’s growth, AI development should not be seen as zero-sum. The U.S. and Chinese AI ecosystems have benefited heavily from collaboration and competition, and keeping the global AI ecosystem open and vibrant will be the best result for both countries. This does not mean that the United States should sit back and ignore Chinese efforts to steal intellectual property and acquire unique technologies with potential national security applications from US companies. Instead, the United States should be cognizant of the cost of these efforts to the U.S. innovation base and we should be clear about our concerns. That China is growing its innovative base and its contributions to the global AI ecosystem is not the problem: the problem is that China has often engaged in illegal and predatory practices to strengthen its domestic AI industry at the expense of U.S. firms and global innovation. The U.S. response should reflect this. One of the United States’ greatest advantages over China is that China is pouring its elite talent and investment capital into U.S. companies and U.S. institutions. The United States should protect the products of U.S. innovation, but that does not mean it should send these resources back to China to support the growth of China’s domestic AI industry at U.S. expense.

Even more important than combating malicious and illegal practices by Chinese firms, the United States must focus on building its own domestic AI innovation capacity. While China is focusing on building its domestic innovation capacity and shaping domestic incentives to support AI development, U.S. policymakers are instead focused on cutting off China’s access to U.S. innovation. China’s ecosystem approach is based on American ideas. NGAIDP draws heavily on concepts developed in two 2016 reports developed by the Obama administration outlining the key factors needed to support continuing U.S. leadership in AI. But while China has invested significant resources in implementing these ideas, the U.S. government’s commitment has been largely rhetorical.

The Trump administration’s February Executive Order on Maintaining American Leadership in Artificial Intelligence emphasizes the right themes, highlighting the importance of building an AI talent pool, expanding AI R&D, developing enabling technologies and infrastructure, and engaging with international partners. But little has been done to implement the executive order, and in fact, many of the Trump administration’s policies run completely contrary to the order’s stated objectives, for example, by slashing budgets at federal agencies that support scientific research, cracking down on visas for people with advanced education and skills, and slashing the U.S. education budget.
In order to maintain U.S. leadership, the United States must take a page out of China’s book. China is investing significant financial, human, and political capital into its drive for AI leadership, taking a strategic approach that brings together policymakers, business leaders, technologists, and the Chinese public to strengthen domestic innovation and build China’s AI ecosystem. China remains behind the United States in fundamental research and key AI technologies, but unless the United States puts its money where its mouth is, China will continue to catch up. Whoever “leads” in the “AI race” or “dominates AI,” China’s AI industries are growing and innovating, and the Chinese government’s strategic approach will continue to drive growth in its AI industry and its contributions to the global AI ecosystem – for better and for worse.

To read the full report, click here