The development of technology, while intended as a mechanism of ease and increased output, has oft been used as a political catalyst. Take for example the industrial revolution, intended as a cost-effective mechanism to ease the production of goods, but inevitably shaped by greed to fuel the capitalist machine that forms the basis of modern society today. This example, while more general and encompassing in nature, is not in isolation. Current governments are no strangers to weaponising technology to exert political influence. The global permeation of AI has caused concerns over its usage in tactics of modern warfare. Sam Altman of OpenAI said that he worries about “authoritarian governments developing this,” in reference to AI development and weaponisation. AI is increasingly being integrated into weapons of authoritarian control. China’s different AI instrumentalisation strategies speak to this larger pattern. China has set an investment goal of 1.1 billion yuan for building agentic AI to be utilised for the spread of Xi Jinping Thought propaganda on his classic Socialism with Chinese Characteristics. Geedge Networks, a Chinese-origin company, is attempting to train AI to predict potential political dissenters. AI is already used as a surveillance tool in areas such as Xinjiang to simplify the identification of different ethnic groups for the purpose of suppressing them and committing a number of human rights violations. These exemplify the tilt towards AI weaponisation to fulfill political goals that are often authoritarian in nature.
‘Xinhua Yudian,’ the agentic AI that is to spread Xi Jinping Thought, is the brainchild of China’s centralised media house, Xinhuanet. Its purpose is to disseminate “Xi Jinping Thought,” the line of political consideration developed by Xi in 2012 to maintain “long-term governance” informed by Marxist ideology. This came as a corollary to Deng Xiaoping’s ‘socialism with chinese characteristics,’ which amalgamated aspects of capitalist economies with heavy-handed government control. Xi’s version of this framework highlights the need to quell corrupt forces in order to ensure the smooth functioning of the state machine. This is evidenced by Xi’s anti-corruption purge of high-ranking party officials. While the justification of effectively removing corruption from the party cadres appears sound, the process takes a more authoritarian form, seeking to “sideline rivals and consolidate power,” as Foreign Affairs magazine reporters Neil Thomas and Shengyu Wang wrote last month. This strategy employs political psychology and translates it into media propaganda to influence the Chinese layman. This propaganda strategy, which has traditionally employed offline media and later social media, is expanding to the realm of agentic AI with Xinhua Yudian. This multi-faceted approach hammers the state’s political thought deeper into the crevices of everyday society, emphasising authoritarian control through ideological conformity and narrative command through Xi Jinping Thought.
Geedge Networks’, a Chinese company involved in producing a surveillance software used to monitor online activity for the Chinese state, is reportedly designing a more advanced AI software that will be used to pinpoint political dissidents based off their online activity. In addition to this, the software reportedly may be able to predict future dissenters through their online language patterns and other activity. Such a software would place a great deal of information in the hands of the government, violating not just basic considerations of personal privacy, but also running parallel to the ‘anti-corruption’ stance that led to purge of party members by targeting the dissent of common citizens. The AI software will enable increased efficiency in combing through common citizens’ online activity. AI softwares are also not error-free, leaving room for false accusations without the right checks and balances. The case of the political consequences of assessing and classifying dissenters is not one in which errors can be tolerated – the effects of such circumstances may be disastrous.
China’s surveillance strategy also extends into electronic surveillance to identify individuals that remained in contradiction to the CCP’s goals for the country. These goals include the ethnic homogenisation of the population, especially in border regions such as Xinjiang and Tibet. Electronic surveillance in these areas has now incorporated AI to track the facial features of individuals and classify them according to ethnicity, in a gross violation of human rights. AI surveillance in these areas has almost taken on the form of a panopticon, where the state’s authoritarianism has solidified into a digital prison.
The different manners in which China has incorporated AI to propagate its ideology and surveil its citizens is a warning sign for loosening democracies worldwide. The spread of propaganda through agentic AI causes a stronghold on narrative control, enabling mass psychological control. The integration of AI into surveillance softwares holds massive implications for wrongful convictions, especially in the case of predicting dissenters. The online surveillance software is being exported to countries such as “Ethiopia, Kazakhstan, Myanmar and Pakistan,” leaving room for potentially similar exports to countries that are finding it difficult to maintain democratic ideals. It would add insult to injury to ignore these examples bringing attention to the sheer potential of increasing digital authoritarian control, to say the least.
