Welcome back to The AI China Report. This week we’re covering new reports on Chinese intelligence agencies using AI for espionage, the latest on the Nvidia-China chip saga, and criticisms from Capitol Hill on Microsoft’s relationship with China.
QUOTE OF THE WEEK
“We’ve been counting tanks and understanding the capability of missiles for longer than we have been as sharply focused on the capability of semiconductors or A.I. algorithms or biotech equipment.”
- C.I.A. Deputy Director David Cohen on the important of U.S. intelligence to focus on Chinese AI capabilities for national defense
ONE BIG STORY: AI IS BOLSTERING CHINESE ESPIONAGE
In July 2018, Xiaolang Zhang, a former Apple employee, was arrested at San Jose airport en route to Beijing. He was charged with stealing trade secrets from Apple's autonomous-driving program. The case of Zhang, who pleaded guilty to stealing trade secrets from Apple, is just a small part of a much larger picture. U.S. authorities suspect that Chinese intelligence is using stolen data from various sources, including the Office of Personnel Management and health insurers, to track U.S. spies and officials. Recent warnings from U.S. intelligence suggest that China might use AI not just for intellectual property theft but also for large-scale data collection on Americans. F.B.I. Director Christopher Wray has expressed concerns about AI amplifying China's hacking operations, using American technology against the U.S.
Over the past decade, Beijing has been linked to massive data hacks, including those at Marriott International, Equifax, and Anthem, among others. These thefts, involving personal information of millions of Americans, have raised alarms about China's capability to use AI to analyze and exploit these vast data sets.
The situation is complex, with China denying involvement in hacking U.S. networks. The Chinese Foreign Ministry has countered U.S. accusations by labeling the U.S. as the "biggest hacking empire." Chinese intelligence officials themselves even admire the U.S. for its digital espionage programs, openly discussing which tactics they could mimic and implement to improve their own programs.
Recently, The Ministry of State Security (M.S.S.), China’s principal intelligence agency, which has both domestic and foreign intelligence responsibilities (think of it as the F.B.I. and C.I.A. combined), has been using AI to develop comprehensive profiles on individuals of interest in Beijing's embassy district. They analyze behavior patterns and vulnerabilities by integrating data from various sources, including surveillance cameras, car license plates, and cellphone data.
This shift towards advanced technology marks a significant transformation for the M.S.S., which historically relied on more traditional methods of information gathering. The agency, once known for gathering intel through social interactions at embassy events, is now at the forefront of technology-driven espionage, rivaling the C.I.A. in the U.S.
The U.S., recognizing the strategic importance of these technologies, has doubled its spending on intelligence activities in China under the Biden administration. The C.I.A. has shifted its focus to gathering information on Chinese companies involved in developing cutting-edge technologies, a change from its previous focus on more traditional defense intelligence.
Despite the technological advancements, the M.S.S. maintains a comprehensive approach to intelligence, combining foreign intelligence gathering with domestic surveillance and political control. Its role in defending the Chinese Communist Party against perceived threats is a top priority. The agency has adapted to the digital age, utilizing social media and other platforms to disseminate its messages and reinforce its objectives.
The rivalry between the U.S. and China's intelligence agencies echoes the Cold War-era competition between the C.I.A. and the K.G.B. However, the current context is distinct due to China's economic growth and its ability to harness emerging technologies like AI, which the Soviet Union could not match during the Cold War.
This technological arms race has significant implications for global security and espionage. As AI and other advanced technologies become more integral to intelligence operations, the nature of espionage is transforming, leading to new challenges and requiring innovative approaches from intelligence agencies worldwide.
NEWS ROUNDUP
China imposed sanctions on the U.S. data analytics firm Kharon and two analysts for their work monitoring human rights abuses in Xinjiang. (WSJ)
Baidu's AI chatbot, ERNIE Bot, has reached over 100 million users four months after regulatory approval for mass launch, as announced by the company's CTO, Wang Haifeng. Although active user numbers were not disclosed, the bot has produced 3.7 billion words in workplace contexts and written 300 million lines of code. (Tech Node)
Nvidia is set to release a modified version of its RTX 4090 gaming processor in China, featuring 11% fewer CUDA cores, to comply with U.S. export restrictions that limit China's access to advanced chips. (CNBC)
U.S. lawmakers are expressing concerns over Microsoft's engagement with China in developing AI technology, particularly after the company's president, Brad Smith, expressed a willingness to assist in China's digital transformation. These ties, highlighted during a recent meeting with China's Minister of Commerce, are seen as a potential national security risk by critics, including Senator Josh Hawley, who advocates for congressional intervention to prevent such partnerships. (NY Post)
China has officially assessed and approved the first group of four large generative AI models, developed by 360 Group, Baidu, Tencent, and Alibaba, establishing a compliance directory for the AI generated content industry. This move, part of a government-led standardization initiative, aims to foster regulated growth in the sector, with these models showing advancements in capabilities and security (Global Times)
DATA TO USE
Know someone who should be in the know about AI and China? Share The AI China Brief 🇨🇳 with your friends and colleagues!
Have feedback on how we can improve? Want to discuss this issue’s content? We’d love to hear from you in comments below!