A Chinese security officer wears a protective mask at the end of the closing session of the National People’s Congress at the Great Hall of the People on May 28, 2020 in Beijing, China. (Photo by Kevin Frayer / Getty Images)
While there is a heated and ongoing debate about who is ahead in the race for artificial intelligence, it is clear that the United States and China, in addition to the military around the world, see the advantage. critical that technology could offer in the event of a conflict. In the editorial below, Georgetown researcher Ryan Fedasiuk explains what he and his colleagues have discovered about China’s AI surge in public records, and how it might reveal some strategic weaknesses. .
For the first time in history, earlier this year, an artificial intelligence system reportedly defeated one of the People’s Liberation Army (PLA) top fighter pilots in mock air combat. Chinese state media hailed the achievement as a milestone in the country’s military modernization. But almost as important as the event itself was the fact that it happened just a few months after the U.S. military reached the same milestone.
For years, pundits have written about China’s plan to use AI for a battlefield advantage, but have cited U.S. advantages in hardware and manpower development as sustainable sources of American strength.
As tensions mount between the United States and China, and some experts warn of a looming crisis in Taiwan, U.S. policymakers and defense planners should understand the types of AI systems already available for it. Chinese army and take measures to defend the advantage of the United States.
In a new report for the Center for Security and Emerging Technology (CSET) at Georgetown University, my coauthors and I sifted through 21,000 equipment contracts issued by PLA units and companies. state-owned defense in 2020. Almost 350 records in our dataset were related to AI systems. and equipment, providing an extraordinarily detailed and fully open-source view of China’s efforts to build a “smart†force.
AI is the foundation of the PLA’s mission to become a “world class” army capable of competing with the United States. First unveiled in 2017, Beijing’s next-generation AI development plan set China’s goal of becoming the ‘world leader’ in AI by 2030 – and it’s clear that this objective extends to military affairs.
Indeed, many AI projects identified in our study explicitly focus on degrading and countering the systems at the heart of the U.S. Army’s concept of joint combat, using techniques such as adaptive radar jamming and vulnerability blurring. Research papers and textbooks from Chinese defense universities even discuss the use of machine learning systems to counter specific US drone swarm projects like Locust and Gremlins.
In addition, the PLA supports its ambitious goals of developing AI with significant investments. Despite the difference of several hundred billion dollars between the budgets of the US and Chinese military, we estimate that the two countries invest roughly the same amount in AI for military use – a few billion dollars every year. Between April and December 2020, more than one in 20 public contracts passed by the main service branches of the PLA was related to AI or “smart†equipment. These projects include all kinds of autonomous vehicles, surveillance systems, training simulators and battlefield decision support software.
In particular, the PLA is investing in AI capabilities designed to scramble, blind, and hack C4ISR systems that link US assets together. Dozens of Chinese military contracts in our study relate to AI systems used in cognitive electronic warfare. Throughout 2020, PLA units and state-supported research institutes have also awarded contracts for “microwave reconnaissance jammer drones” and “electromagnetic weapon” payloads that can be attached to swarms of small unmanned aerial vehicles (UAVs) and flown into enemy airspace. A handbook attributed to Chinese electronic warfare operators compares the US military’s battlefield information networks to the human nervous system: “Once the ‘tendons and veins’ or ‘blood vessels’ are cut off, people will be paralyzed or even killed. “
Additionally, the PLA Navy is hoping the AI ​​will compensate for its long-standing drawbacks in submarine warfare. Since 2015, Chinese research institutes have launched hundreds of autonomous underwater vehicle research programs and have made steady advancements in fundamental technologies such as battery life and high seas communications. a potential crisis in Taiwan, AI-based systems could extend the PLA’s navy underwater reconnaissance operations far beyond the first island chain, and cheap autonomous platforms could be useful for mine laying or anti-submarine warfare.
Despite the PLA’s significant progress in adopting AI-enabled systems, however, there are at least two clear vulnerabilities in its plan to create an “intelligent” force.
First, while Chinese military leaders plan to exploit weaknesses in America’s sensor and communications networks, it’s unclear how they plan to build their own resilient cloud-based networks. PLA officers often write that the US military is prone to information manipulation and data poisoning, even calling data integrity the “Achilles heel” of common command strategy and control all areas of the United States.
But in a potential conflict, the PLA itself would also struggle to ensure the integrity of the data used to train its own AI systems, let alone the inherent fragility of computer vision and recognition systems. AI-based objects. None of the 350 unclassified Chinese military contracts in our study focus on building resilient networks or secure datasets.
To mitigate the threat posed by Chinese military AI systems, US defense planners should increase investment in research into counter-autonomy and conflicting AI that exploits vulnerabilities in Chinese systems, while enhancing the robustness of the systems. American AI systems.
Second, China’s “smartness†strategy relies entirely on access to AI chips designed by US companies and manufactured in Taiwan and South Korea. The supply of this high-end microelectronics is however far from guaranteed. In fact, the United States and its allies have already taken several measures to deprive Chinese military companies of the chips needed to train advanced machine learning models.
But to effectively slow China’s military progress in AI, U.S. policymakers should continue to increase investment in organizations that are supposed to regulate technology releases, like the Commerce Department’s Office of Export Law Enforcement. ; and crack down on third-party middlemen who supply the Chinese military and defense industry with US-made equipment.
While US leaders shouldn’t dismiss the Chinese military’s AI progress, there are clear vulnerabilities in its “smart†war plan. By striking the appropriate balance between promoting innovation at home and preventing the leakage of American technology abroad, the United States can maintain its lead in military AI.
Ryan Fedasiuk (@RyanFedasiuk) is a research analyst at the Center for Security and Emerging Technologies at Georgetown University (@CSETGeorgetown), focusing on military applications of AI and China’s efforts to acquire foreign technology .
