9 June 2020

CHINESE DEBATES ON THE MILITARY UTILITY OF ARTIFICIAL INTELLIGENCE

MICHAEL DAHM

The Chinese military believes it is losing a high-stakes competition with the United States and Russia to lead the world in artificial intelligence (AI). In articles like, “The Quiet Rise of an Artificial Intelligence Arms Race” (人工智能军备竞赛正在悄然兴起), Chinese military authors point to a quote from Russian President Vladimir Putin, that whoever leads in AI will “rule the world.” As evidence of the U.S. military’s ambition to dominate in this field, they cite findings about AI in future warfare from the U.S. National Security Commission on Artificial Intelligence, calls for the United States to ally with other nations against Chinese AI development, the Department of Defense AI strategy, and the establishment of the Pentagon’s Joint Artificial Intelligence Center. In 2017, China was among the first nations to advance a national-level AI development strategy that broadly addressed AI’s role in economic development.

The Chinese military, however, has been opaque about its AI strategy and intentions. Undoubtedly, Chinese military officials understand they must compete with the United States by adapting quickly to changes in warfare brought about by AI and autonomous systems. An examination of the ongoing debate within the ranks of China’s People’s Liberation Army (PLA) about the transformation of warfare by AI — what they call “intelligentized warfare” (智能化作战) — reveals that this new form of warfare is an extension of existing Chinese strategy and operational concepts.


AI may allow China to realize its long-standing, information-centric military strategies. The American military tends to focus on how AI can enable lethal attacks against opposing forces. Chinese strategists tend to argue that AI technologies should be used kinetically and non-kinetically to dominate information systems and networks, to effectively paralyze an opponent’s joint force. Information warfare and information control are at the heart of the PLA’s approach to warfare — and AI. Countering China’s strategy will require a defensive and offensive use of new AI technologies. In a future confrontation, the U.S. military will need to employ AI and autonomous capabilities to enable and defend its information system-of-systems while simultaneously using AI technologies to attack China’s information-centric strategy and capabilities.

China’s Established Military Strategy Will Be Enhanced by AI

The PLA’s overarching strategy for defeating the U.S. military, or any foreign adversary, is to dominate in a system-of-systems confrontation. This method of warfighting focuses on creating disruption or paralysis across an enemy system-of-systems versus emphasizing the attrition of forces. First, the PLA will attempt to crash the adversary’s information networks using kinetic and non-kinetic means. The Chinese military believes that information is the critical element that binds and enables a larger military system-of-systems. Second, the PLA intends to eliminate individual elements of a now-disaggregated enemy force with long-range precision fires. This Chinese military doctrine has been described as “systems confrontation,” but that short-hand does not accurately capture the potential for a cascade of compounding effects within a complex system-of-systems and the resulting paralyzing outcomes. AI may provide a critical means to that end.

American assessments of military AI often focus on the second step — coordinated lethal attacks using autonomous systems against opposing forces. Drones and other autonomous systems are certainly under development in the PLA. However, the Chinese focus is currently on developing AI technology, methods, and tactics to precisely target key elements within an enemy system-of-systems. The objective is to paralyze the adversary, and goes well beyond merely “throwing sand in the gears” of the enemy joint force. If successful, the large-scale attrition of forces may not even be necessary.

The use of AI in a system-of-systems confrontation conforms with and enables existing Chinese military doctrine on informationized warfare. The PLA believes that the center of gravity in modern military operations has shifted from concentrations of forces to information systems-of-systems — everything from target detection to communication to information processing to command of action. Modern military information systems-of-systems are vast, complex, and in the future will likely be managed by AI. Therefore, it follows that they can only be analyzed in real-time and attacked using AI.

The PLA’s objective is to use AI algorithms, machine learning, human-machine teaming, and autonomous systems collaboratively to paralyze its adversaries. The ultimate goal for the Chinese military appears to be cognitive advantage — the ability to adapt one’s system-of-systems faster than one’s adversary. The Chinese seek to use AI to deliver precise effects to immobilize their adversary while defending their own system-of-systems. Any Chinese military challenger would be wise to understand the implications of how future AI capabilities may be employed to realize Chinese goals in system-of-systems confrontation.

An Intelligentized “Form of War”

The English-language version of China’s 2019 Defense White Paper observes a change in modern warfare: “War is evolving in form towards informationized warfare, and intelligent warfare is on the horizon.” A translation of the Chinese-language version, however, reveals that the change is not about moving toward informationized warfare, it is about an evolution in informationized warfare: “The form of war is accelerating toward an informationized warfare evolution, there are indications intellgentized warfare is emerging” (战争形态加速向信息化战争演变,智能化战争初现端倪). The Chinese “form of war” (战争形态) speaks to the changing character of war; an assessment of the objective basis that will drive the conduct of present and future warfare. The information age had yielded informationized warfare (信息化作战) forming the basis for PLA development since the early-2000s. Chinese military leaders now believe that informationized war is evolving and intelligentized warfare will become the prevailing form of war.

The 2019 forecast of an evolution in the form of war toward intelligentization is very similar to where the PLA found itself in the early 2000s. The 2002 Chinese Defense White Paper stated, “The form of war is developing in the direction of informationization.” What happened in the wake of that top-level proclamation was a vigorous discussion among Chinese military writers parsing out the transformation of warfare by information and modern information technology. That debate and the dialectical back-and-forth ultimately formed the foundations of Chinese informationized warfare theory and doctrine. It bore out informationized warfare “patterns of operations” (作战样式) and “basic guiding thought” (基本指导思想) — informationized warfare theory that placed information control at the center of PLA operational concepts. As with those explorations of informationized warfare in the early 2000s, Chinese military authors are currently discussing the changes that intelligentized warfare will bring to the PLA.

The Chinese military is not a hive mind. Debates within the Chinese military are robust and often contrarian. Western analysts must distinguish between guidance from central leadership and what may be dissent or consensus in a debate. The pages of the PLA’s official newspaper, the PLA Daily (解放军报) and the PLA’s official web site provide a venue for the discussion of emerging phenomenon on any number of topics, including intelligentized warfare. While far from authoritative guidance, the publication of research, opinions, and proposals in official media comes with the tacit endorsement of the Chinese military establishment. The PLA Daily newspaper’s “Military Forum” (军事论坛) offers monographs by select military authors and provides insights into the consensus that may be emerging within the PLA over intelligentized warfare.

Since the Defense White Paper was published in mid-2019, authoritative commentaries have appeared with increasing frequency in official PLA media discussing informationized warfare, intelligentized warfare, unmanned systems, autonomous decision-making, and cognitive warfare. In articles like, “Seize the Commanding Heights of Artificial Intelligence Technology Development” (抢占人工智能技术发展制高点), authors discuss the development of this new type of intelligentized warfare, forecast new types of intelligentized technologies and operational concepts, and ultimately seek to propel PLA thinking ahead of counterparts in the United States. These commentaries demonstrate a trajectory of PLA military thinking that integrates AI with Chinese thinking on informationized warfare and the PLA’s information-centric military strategy.

Six Principles of China’s Intelligentized Warfare

There is a significant amount of overlap among Chinese, Russian, and Western militaries on the topic of military AI. For example, all seem to recognize that autonomous swarms of unmanned platforms may generate advantages in terms of cost, scale, dispersion, and adaptation, enabling lethal saturation attacks. Articles like “Intelligentized Warfare, Where are the Constants?” (智能化战争, 不变在哪里) (January 2020), remind readers that AI may have changed the character of war, but the nature of war endures — war is still a violent action taken to a political end with humans central to the endeavor. Several Chinese authors emphasize that humans will still plan, organize and initiate wars. While man-machine teaming may enhance human cognition and action, Chinese analysts are directed to guard against the “anthropomorphization of weapons and the weaponization of humans” and always place humans in the dominant role.

However, there are a number of unique perspectives emerging from the Chinese debate over AI that may merit attention by Western analysts. Most appear to reflect the premise that AI will enable and evolve existing Chinese warfare theory and doctrine.

First, PLA strategists believe that intelligentized warfare is an evolution of informationized warfare. Several authors point out that intelligentized warfare is essentially highly evolved informationized warfare — system-of-systems confrontation reliant on information moving through digital systems and networks. “How to Integrate the Mechanization, Informationization and Intelligentization of Weapons and Equipment” (武器装备机械化, 信息化, 智能化怎么融) (October 2019) acknowledges that informationization and intelligentization are inextricably linked. But articles like this have begun to categorize intelligentization as an independent, aspirational stage of future military development.

Next, Chinese military authors argue that ubiquitous networks will enable systems-of-systems warfare. In articles such as, “Picturing a New Combat System” (为新型作战体系画个像) (June 2019) they refer to the emergence of “ubiquitous networks” (泛在网络) that will shorten the distance between perception, judgement, decision-making, and action. The Chinese concept of “ubiquitous networks” appears similar to the U.S. Defense Advanced Research Projects Agency’s (DARPA) Mosaic Warfare concept of system-agnostic, flexible, and rapidly configurable networks. The PLA has been focused on military systems-of-systems development for well over 15 years. Ubiquitous networks and AI may give the PLA its own version of mosaic warfare “with Chinese characteristics.”

Chinese strategists also contend that AI enables command and operational design. For example, “How Unmanned Combat can Change the Form of War” (无人作战如何改变战争形态) (August 2019) highlights an underappreciated feature of Chinese military theory — the way in which command and control is exercised through operational design and pre-conflict campaign planning. There are misperceptions that the PLA has a Soviet-style centralized command and control system — the premise being that if PLA conventional forces or autonomous systems are cut off from centralized control, they either will be unable to function or will make haphazard decisions. However, the PLA believes that command and control can be “built-in” to system design and operational plans to mitigate threats of either human or machine errors in combat. Consistent with its Marxist-Leninist intellectual heritage, the PLA conceptualizes war as a scientific process that can be deconstructed, allowing for calculated outcomes. AI and machine-learning may provide the Chinese military with the algorithms and tools it believes it needs to determine end results and develop invulnerable systems-of-systems, operational capabilities and military plans.

Fourth, there is an expectation that AI will enable new operational concepts. “Intelligentized Warfare, Where are the Changes?” (智能化战争, 变化在哪里) (January 2020) is the companion piece to “Intelligentized Warfare, Where are the Constants?” cited above. In this article, the authors identify several future “patterns of operations” — Chinese operational concepts — including autonomous swarm attrition warfare (自主集群消耗战), autonomous dormant assault warfare (自主潜伏突袭战) autonomous cross-domain mobile warfare (自主跨域机动战) and autonomous cognitive control warfare (自主认知控制战). Swarm attrition warfare capitalizes on decentralized operations and coordinated saturation attacks by large numbers of autonomous, unmanned systems. Dormant assault warfare portends surprise attacks at key points or against critical adversary capabilities by autonomous platforms programmed to lie in wait until activated. Autonomous cross-domain mobile warfare envisions an autonomous force that is highly mobile and can strike on a large-scale and at long-ranges. This last concept presumably borrows from the U.S. Army’s Multi-Domain Operations concept.

Autonomous cognitive control warfare is not explicitly described in the January 2020 article. However, articles like “Cognitive Warfare: Dominating the Intelligence Age” (认知战: 主导智能时代的较量) (March 2020) describe a shift in opposing military centers of gravity toward the cognitive domain. Chinese military authors are fond of invoking the U.S.-originated OODA loop (observe, orient, decide, act). These Chinese authors observe that decision-making is the bottleneck in the OODA loop. Future autonomous systems, they say, will compete for cognitive advantage and thus decision advantage enabling faster cycling of military action to dominate an adversary in “parallel operations” drawing from the U.S.-originated parallel warfare concept.

Fifth, AI offers the PLA the ability to precisely release kinetic energy and paralyze an opponent’s system-of-systems. A Chinese state-media (Xinhua) article, “Military Intelligentization is Profoundly Affecting Future Operations” (军事智能化正深刻影响未来作战) was posted to the Chinese Ministry of National Defense website in September 2019. The article observes that in intelligentized warfare — cognitive-centric warfare — AI and autonomous systems will allow for precise energy release, either dispersed across a system-of-systems or concentrated on a critical node to impose “highly persistent paralysis” on an adversary. Battlefield advantage will go to the force that can dominate the cognitive domain — perceiving, adapting and acting faster than an opponent to impose or reverse system-of-systems paralysis.

Finally, Chinese military analysts recognize that the PLA must accelerate progress on AI. The military scholars cited here uniformly believe that China is well behind the United States in the development of military AI. The emphasis on AI in the Defense Department’s Third-Offset Strategy, or even Russia’s progress in unmanned and autonomous systems, are often cited to illustrate the PLA’s lagging progress, adding a sense of urgency to the need for military AI development in China.

The PLA “Overtaking on the Curve”

AI is a clear priority for the Chinese military. The PLA has adopted a strategy of “overtaking on the curve” (弯道超车), catching up with and passing the United States and Russia by metaphorically turning more tightly in corners as trends in science and technology change direction. There will be myriad innovations that change the trajectory of AI technology in the coming years. Each one will be a potential opportunity for the PLA to close the technology gap with the United States and allow the Chinese military to realize its information-centric military strategy.

According to recent Western studies of the AI industry, the United States appears to be in the lead with China rapidly closing the gap. Assessing military competition in AI is more difficult. The U.S. military’s Fiscal Year 2021 budget proposed $841 million in direct spending on AI (0.1 percent of the $705 billion proposal), but that fails to capture how AI is being integrated throughout different weapons systems’ budgets. Chinese defense spending is even more opaque. However, a significant number of Chinese military institutions do appear to be working diligently on AI innovation.

Assessing how the Chinese and U.S. militaries compare in terms of AI is made even more difficult by the fact that innovation and technical progress on AI has been driven by industry for civil applications. Armed forces will continue to capitalize on the dual-use nature of big data processing and AI algorithms that increase industrial efficiencies and enable commercial autonomous systems.

In the competition to lead in AI, China enjoys the advantage of scale. The Chinese government is accelerating the development of AI technology using entire cities as laboratories. In early April 2020, in the wake of the COVID-19 outbreak, Chinese President Xi Jinping visited Hangzhou and called for Chinese cities to become “smarter.” Hangzhou is just one of several Chinese “smart cities” experimenting with the use of AI for city-wide management and security. Ultimately, the PLA will capitalize heavily on China’s cutting edge progress in civil and commercial AI technology, much of it in cooperation with U.S. and other foreign industries. Accounting for these different development models directed by government or left to private industry may be critical in forecasting outcomes in a military AI arms race.

The Big Picture

The Chinese military, in its own self-assessment, falls well short of a globally present U.S. military that can project an overwhelming joint force anywhere in the world with cutting-edge military technology and a wealth of combat experience. President Xi has mandated that the Chinese military be “fully modernized” by 2035 and a “world-class military” — on par with the United States military — by 2050.

China’s strategy for the use of AI technology is evolving from their interpretation of the character of war and is ultimately an extension of the PLA’s informationized warfare concepts. While there is certainly overlap with U.S. military thinking on the use of AI, Chinese military scholars appear to be reaching different conclusions. U.S. thinking tends to emphasize the role of AI in enhancing firepower- and maneuver-centric strategies. The PLA, on the other hand, is advancing AI concepts that enhance its information-centric military strategies.

New technologies related to artificial intelligence, machine learning, and autonomous systems may provide the PLA with necessary tools to realize its long-standing goals of controlling the information domain, manipulating perception, and paralyzing adversary decision-making. In developing strategies to counter Chinese military capabilities, the Pentagon should pay close attention to the PLA’s evolving warfighting concepts and views on AI in future combat.

No comments: