China uses AI system to test Communist Party officials' loyalty
China deployed an AI system called the Smart Ideological Work Station in 2022 to test Communist Party officials' loyalty to Xi Jinping by analyzing physiological responses and facial expressions during questioning. The system was reportedly developed by the Hefei National Science Center Artificial Intelligence Research Institute and used by the Party's Central Organization Department to screen cadres for political reliability.
What happened
In 2022, the Hefei National Science Center Artificial Intelligence Research Institute, established by the Chinese Academy of Sciences and Anhui authorities following a national laboratory system model, announced on WeChat that it had successfully integrated AI with Party building to assess Communist Party members' acceptance of ideological education. The system, called the Smart Ideological Work Station (ζΊζ §ζζΏε§), used designed questions to test loyalty to Xi Jinping, measuring heart rate, perspiration and facial expressions to determine whether respondents were telling the truth. The announcement was later removed from WeChat.
According to commentary in The Epoch Times, the Central Organization Department of the Chinese Communist Party has deployed AI technology as a method to screen for political disloyalty among officials in recent years, as part of efforts to prevent disloyal individuals from rising to positions of power.
Reported impact
An independent commentator quoted in the article stated that using AI to test loyalty constitutes technological rights abuse, describing the experience as akin to torture for individuals assessed by the system, particularly since most Communist Party members did not join based on genuine ideological belief.
Index analysis
The deployment of loyalty-testing AI reflects broader concerns within the Chinese government about hidden dissent among Party cadres, according to analysis cited in the articles. The system is designed to identify officials with insufficient political loyalty by analyzing physiological and emotional responses during questioning.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.