from Net Politics and Digital and Cyberspace Policy Program

Beijing’s AI Strategy: Old-School Central Planning with a Futuristic Twist

A staff member stands next to robots at a plant of Kuka Robotics in Shanghai, China Pete Sweeney/Reuters

China's new artificial intelligence strategy is a signal that Beijing wants to be a leader in AI. How it gets there is a different story.

August 9, 2017

A staff member stands next to robots at a plant of Kuka Robotics in Shanghai, China Pete Sweeney/Reuters
Blog Post

More on:


Robots and Artificial Intelligence

Lorand Laskai is a research assistant in the Asia studies program at the Council on Foreign Relations. You can follow him @lorandlaskai.

Rarely do government policy documents come off as manifestos about the future. China’s much-anticipated “Next-Generation Artificial Intelligence Development Plan,” which was released last month (see an excellent translation from Rogier Creemers, Graham Webster, Paul Triolo, and Elsa Kania here), is an exception. For a country that releases a bevy of slogan-filled ‘plans’ each year, most of which are as vague as they are forgettable, the AI development plan is different. What the plan lacks in concrete details, it makes up for in vision and ambition. It depicts a future China overcoming the challenges of an aging population and resource constraints through integrating AI into everything from agriculture and manufacturing to governing and public security. Still, the plan also reveals that China’s opening salvo in the global race to dominate AI will rely on old-school centrally guided economic development. The question is, will it work?

In raw numbers, China’s massive ambitions for AI are clear: the plan proposes transforming China’s fledgling AI industry into a ‘world leader’ worth RMB 400 billion ($60 billion) by 2025 and a ‘premier innovation center’ worth RMB 1 trillion by 2030 ($150 billion). With China’s AI industry currently hovering around RMB 10 billion ($1.5 billion), this is a tall order. And even though the plan rattles off a long list of technologies China wants to make breakthroughs in, the plan provides few concrete actionable steps to get there. This is typical for policy from the central government. In China, the central government releases plans or frameworks to communicate to cadres in the bureaucracy and at the local level what the leadership deems a priority. With the AI development plan, top Chinese officials are signaling that the leadership views AI as central to China’s future.

In all likelihood, China’s cadres already knew that. Government ministries, provinces, and large municipalities like Shanghai and Tianjin have already unveiled a slew of policies to nurture the technology as part of the country’s thirteenth five-year plan. Scientists from the Chinese Academy of Science (CAS) and Chinese Academy of Engineering (CAE) have embarked on a series of AI ‘megaprojects’ including ‘China AI 2.0’ and ‘China Brain Project,’ which some have compared to the U.S. Apollo program. China’s commercial enterprises are responding, if not already leading the way. The big three—Baidu, Alibaba, and Tencent—are investing heavily in AI. At the same time, a new class of large Chinese AI mega-enterprises, like Sensetime, Mobvoi, and iFlytek, are rising. Unconstrained by strict privacy norms, China is arguably commercializing certain AI applications, like facial and object recognition, at a faster pace than the West.

What’s the point of the central government’s plan if major actors are already making AI a priority? In one sense, the plan is China’s Ministry of Science and Technology (MoST) attempt to reestablish its prerogative over high-tech developments. Rather than allowing other ministries or experts at CAE and CAS to steer the way on China’s AI research agenda, the bureaucrats at MoST are reasserting their voice in the process through the new ‘AI Plan Implementation Office.’ This can be bad for innovation and the overall development of the technology, since it will insert an additional layer of bureaucracy and competing objectives to the mix. Top MoST officials already have a history of intervening into research programs and funding decisions to meet short-term political objectives.

Viewed more broadly, China’s AI development plan can be accurately framed as a roadmap for how the central leadership wants to shape and spread the inchoate technology. This is why the plan is packed with fantastical societal visions of the technology’s applications. These include AI enhanced policymaking, ‘smart courts’ that streamline everything from evidence collection to trial, and some applications that defy easy description, like using AI and blockchain to “enhance social interactions and promot[e] credible communication.” One of the more interesting policies embedded in the plan is the creation of ‘AI innovation application pilot zone’—essentially, real world labs for pioneering new applications for AI. These zones could support experimentation, providing AI scientists with all the data and legal leeway they need to innovate new applications for AI, especially if the application fits the government’s interpretation of public good.

Less sensational–and probably more realistic and consequential–steps are outlined in the plan as well: creating industrial parks to nurture A.I. companies, pushing companies in all sectors of the economy to upgrade their ‘smartness levels’ through integrating A.I. into core operations, and reorienting the education system to train Chinese students to work with A.I. These are hallmarks of China’s central-guided development, and in many ways the main advantage that the Chinese government is betting on to win the A.I. race. That is, through a central push, China believes it can launch a coordinated strategy, leveraging more resources, launching more moonshot projects, and think bigger than any other government can.

Will Beijing’s plan work? Like many grandiose ambitions of central planning, the more spectacular elements will likely fail. Still, the plan might still work if only because China AI plan commits resources to a society-wide push for AI. Former Baidu Chief Scientist Andrew Ng’s comparison of AI to the invention of electricity is instructive. Like electricity, the transformative effect of AI will be immediate and impact every aspects of the economy and society. But as Miles Brundage and Joanna Bryson point out, electrification was also multiple decade process (the U.S. did not fully electrify until 1950) and required the mobilization of considerable state resources. AI will also demand a concerted push if the technology is to spread throughout society. With its AI plan, Beijing is signaling that China is determined to being the first country to make this push.

Creative Commons
Creative Commons: Some rights reserved.
This work is licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) License.
View License Detail