Artificial intelligence (AI) is a key technologies of the future – and not just because of civilian applications like ecommerce, self-driving cars, and online search and personal assistants. AI will transform militaries through innovations in intelligence, surveillance, reconnaissance, logistics, command and control capabilities, weapons systems, and so on. Last year a landmark U.S. National Security Commission on Artificial Intelligence (NSCAI) report found, “The ability of a machine to perceive, evaluate, and act more quickly and accurately than a human represents a competitive advantage in any field—civilian or military.” In other words, national military superiority will depend on which nation can best develop and deploy AI, notes a new report from think tank China Tech Threat.
Right now the People’s Republic of China (PRC) is racing to achieve (if not sustain) AI competitive advantage over the United States. The Center for Security and Emerging Technologies (CSET) study found that of 343 military contracts with AI firms, the PRC is most focused on procuring AI for military applications. The NSCAI likewise affirms the PRC’s intentions: “China’s plans, resources, and progress should concern all Americans. It is an AI peer in many areas and an AI leader in some applications. We take seriously China’s ambition to surpass the United States as the world’s AI leader within a decade.”
However the PRC is already known for its use of AI for civilian repression. IPVM exposed the chilling aims of the PRC’s AI-automated racism to surveil Uighur Muslims. Its reporting has been corroborated and published jointly with the Washington Post, NYTimes, and BBC. Huawei partnered with leading PRC AI/facial recognition developer Megvii to patent the so-called “Uighur alarm” to identify Uighurs by face and track their movement, turning the Xinjiang province in a de-facto “open-air prison” for 25 million people. The Department of Commerce sanctioned Megvii in 2019 for its flagship product Face++.
PRC computer giant Lenovo, needing facial recognition tech for its laptops, was a lead investor in the now-restricted Face++ and co-founded the FIDO (Fast ID Online) Alliance in Mountain View in 2012 Lenovo also develops proprietary facial recognition technologies like LeFace and Veriface, which are purported to be even more “racially” accurate than Face++.
PRC embeds AI and facial recognition tech its countless devices and systems for export, thus enabling a global PRC dragnet. Many authoritarian nations purchase this tech for specific political, repressive purposes. For example, some 300 Huawei surveillance cameras in enable repression in Myanmar’s authoritarian regime. While the Soviets gave away AK-47s get new members of its bloc, the PRC sells AI-enabled tech to build police states in emerging countries.
Unsurprisingly China has shrewdly enhanced its AI capabilities with American tech The Department of Defense’s Defense Innovation Unit tracked at least 81 instances of PRC investment in American AI companies from 2010-2017 and suggests a high probably that American AI is already deployed in the PRC’s military. The As CSET reports that just 8 of the of the 273 PRC military suppliers t are named in U.S. export control and sanctions regimes. Essentially there are little to no controls on tech from some 7000 American AI companies making their way to the Chinese military, wittingly or not. As China Tech Threat detailed, Congress instructed the BIS to identify and detail AI technology which should be export-controlled for military application, but the agency has not moved on this task in almost 4 years.
Fortunately BIS has new Undersecretary who has pledged to revisit the security of technological export to China. Unique for his defense and supply chain background, Alan Estevez should leverage his Pentagon relationships and expertise to fast-track a BIS strategy for AI. This can include creating a cadre of personnel with military AI expertise, including individuals who can work part-time from the civilian sector. Similarly, the NSCAI report proposes the creation of a “Digital Corps” workforce of AI-skilled professionals who are assigned to different agencies at different times to work on key problems and projects. Additionally, BIS can educate U.S. firms about the dual-use applications of AI and how it can be weaponized by adversaries. Indeed, it will take a whole of nation effort to win what will likely be the critical battlefront of the 21st century.