Chinese AI Project Scouts Spies in Taiwan Chip Industry
The Chinese government is employing an AI-driven Big Data mining program to monitor ethnic Chinese individuals abroad, particularly targeting those in the semiconductor industry. This initiative reportedly aids in designing influence operations aimed at high-ranking figures within the artificial intelligence (AI) sector.
As the global AI landscape evolves, nations are exploring ways to leverage this technology to enhance their power. While the United States focuses on immediate AI applications, such as tracking immigrants and political protestors, the Chinese Communist Party (CCP) has established a more extensive data center. This facility facilitates massive web scraping from social media and other public sources, utilizing machine learning algorithms to generate actionable intelligence for officials in Beijing.
Documents leaked to the watchdog group NetAskari reveal the dual nature of this program, which combines public relations efforts with surveillance tactics. Some data collection efforts resemble targeted propaganda campaigns akin to those used by Cambridge Analytica. In more troubling instances, the program identifies individuals in politically and economically significant roles who could potentially serve as spies or unwitting assets.
According to a NetAskari analyst, “A foreign nation with rather undemocratic tendencies is running an extensive, AI-charged analysis and information-scraping operation to find ways to influence other nations’ internal social, economic, and political landscapes.” The analyst emphasized the shocking nature of targeting the Chinese diaspora for these operations.
Targeting Semiconductor Industry Leaders
On November 8, NVIDIA CEO Jensen Huang disclosed an affair with a spy linked to the Ministry of State Security (MSS). The agent, Alice Huang, was involved in redirecting NVIDIA semiconductors to China, where they are prohibited, and relayed intelligence from Huang to the CCP.
Meanwhile, the United Front Work Department (UFWD) has taken a less overt approach to infiltrate Taiwan’s semiconductor industry. The UFWD’s operations focus on shaping media narratives about China and engaging with Chinese diaspora communities, sometimes to their detriment. In this context, the UFWD has been investigating ethnic Chinese individuals in the semiconductor sector.
A leaked report titled “Profiles of Ethnic-Chinese Leading Talent in the Chip Field, Their Circle Characteristics, and Work Recommendations” highlights executives from the Taiwanese AI company Skymizer, top-performing students from a leading Taiwanese research university, and a director from an AI research center at Microsoft. The report outlines strategies for building relationships with these influential individuals, such as leveraging non-official organizations like hometown associations and professional societies.
Overview of the Big Data Platform
The system generating these insights is known as the “Overseas Chinese Affairs Big Data Platform,” housed in a 500-square-meter laboratory at Jinan University in Guangzhou. This facility is designed to process vast amounts of online information about overseas Chinese communities, identify key individuals and relationships, and produce actionable reports for CCP leadership.
The platform scrapes data from various social media sites, including Instagram, Facebook, and Chinese platforms like Weibo and WeChat. The data collection is conducted using machines located in foreign countries, often obscured by rotating proxy networks to evade detection. According to leaked documents, this strategy aims to ensure the security of data collection and prevent tracing back to the collection system.
Once collected, the data is sent to China for processing, where it is translated into Chinese, keywords are extracted, and connections are established. The UFWD cross-references this information with existing records of notable individuals in media, science, and politics, as well as those on blacklists. This process feeds into a “business system” that identifies public sentiment and the outcomes of influence operations.
AI’s Role in Expanding Influence
The NetAskari analyst describes the Overseas Chinese Affairs Big Data Platform as a global surveillance tool that utilizes open-source intelligence (OSINT) and Big Data analysis to identify potential collaborators and partners for the UFWD. This effort aims to extend the Communist Party’s influence abroad and foster a “China-friendly” environment in other nations.
While the UFWD has conducted similar operations in the past, the analyst notes that AI and modern technology have significantly accelerated these efforts. The leaked documents reference outdated machine learning and natural language processing algorithms, leaving unclear how contemporary large language models may enhance the system’s capabilities.
The analyst expresses concern over the implications of such surveillance, stating, “A surveillance state can’t accept that its own citizens are escaping to other nations, outside the reach of Beijing’s grasp.” With advancements in remote data collection methods, traditional barriers to expanding influence no longer apply, raising alarms for citizens in the US and Europe, even those without direct ties to China.
Source: Dark Reading