DIGITAL LIFE

China is increasing its use of artificial intelligence to strengthen social control mechanisms, expanding the reach of censorship and surveillance over the population

This conclusion is part of a new report from the Australian Strategic Policy Institute (ASPI), an Australian institution that details how the Beijing government has integrated cutting-edge technologies into the country's digital monitoring system and judicial apparatus.

According to the study, artificial intelligence is being used to make the tracking and suppression of politically sensitive content more efficient. This practice, already a hallmark of the Chinese censorship apparatus, gains speed and precision with systems capable of scanning large volumes of data, identifying keywords, and reducing the dissemination of messages critical of the government or leader Xi Jinping.

Nathan Attrill, senior analyst at ASPI and co-author of the research, told the Washington Post that the technology does not inaugurate a new model of repression, but intensifies already established methods. According to him, AI allows the Chinese Communist Party to monitor "more people, more closely, and with less effort," deepening control patterns previously executed mostly by human teams.

The technological dispute with the United States also influences this scenario. While the US and China compete for global leadership in artificial intelligence, Beijing is expanding the domestic use of the technology, relying on the collaboration of large national companies such as Baidu, Tencent, and ByteDance. These companies receive access to immense sets of government data, which accelerates the development of more advanced models.

The report highlights that the companies act as "assistant sheriffs," responsible for moderating content that goes beyond the scope normally adopted by platforms in the West. While social networks in other countries only remove illegal material, such as pornography, Chinese firms also need to eliminate content that could irritate the central government. Tools such as Tencent's content security audit system, for example, assign risk scores to users and track repeat offenders across various platforms.

This surveillance industry has also become a business. Companies like Baidu market automated moderation systems to other companies, expanding the reach of censorship and, according to researchers, "storing market principles in service of authoritarianism." Despite increasing automation, the adopted model is hybrid: human teams remain essential to interpret political nuances, identify codes used to circumvent supervision, and compensate for technical flaws.

Surveillance is even more intense over ethnic minorities, such as Uyghurs and Tibetans, targets of expanded monitoring in recent years. Because language barriers hindered tracking, the government is investing in the development of language models specific to regional languages. A laboratory created in Beijing in 2023 works with languages ​​such as Mongolian, Tibetan, and Uyghur, with the aim of analyzing public opinion and promoting what the government calls "ethnic unity."

The report also details how AI is being incorporated into the criminal justice system. Technology is already appearing in the identification of suspects at protests through facial recognition, in the screening of judicial documents by "smart courts," and even in prisons capable of predicting the emotional state of detainees. Researchers who had access to one of these systems in Shanghai warned that the tools could compromise judicial impartiality by introducing multiple "black boxes" that are impossible to audit.

Experts point out that the use of AI in the Chinese justice system has gone through distinct phases: it began with enthusiasm and exaggerated expectations, followed by a period of caution, and is now experiencing a stage of reflection on limitations and risks. The accelerated adoption of the technology, encouraged by guidelines from the central government, often leads local authorities and companies to exaggerate its capabilities to obtain contracts, making it difficult to measure the real impact of these systems.

The report concludes that, despite advances and efficiency in some processes, the expansion of AI in China raises profound concerns about privacy, transparency, and discrimination. For researchers, the lack of clarity about how the models work and the risk of inherent bias make the ecosystem particularly dangerous, even more so because Chinese companies have global ambitions and export these systems to other countries.

Key developments include:

Accelerated Monitoring: AI allows the Chinese government to scan and analyze vast volumes of digital content in real time, identifying and suppressing politically sensitive material much faster than manual methods.

Predictive Control: Authorities are using algorithms to analyze patterns of online behavior and sentiment, aiming to anticipate and neutralize dissent or protests before they occur, which experts describe as "preventive repression."

Minority Surveillance: Reports indicate that the government is developing specific AI tools to deepen the monitoring of ethnic minorities, such as Uyghurs and Tibetans, including through language models in their native languages, both inside and outside China.

Integration into the Judicial System: AI is being implemented in courts and prisons to assist in processes, from drafting documents to recommending verdicts and sentences, raising serious questions about impartiality and accountability.

Multimodal Censorship: In addition to text, new Chinese AI systems are capable of censoring politically sensitive images and videos, adding a new layer to the country's "Great Firewall."

Collaboration with Technology Companies: Large Chinese technology companies, such as Tencent, Baidu, and ByteDance, are developing and selling AI-based censorship platforms to other organizations, creating a domestic market for these control tools.

These actions have led to discussions and restrictions on the use of Chinese technologies in other countries, such as the United States and the European Union, due to concerns about privacy and alignment with the values ​​of the Chinese Communist Party. China, in turn, advocates for the creation of a global organization for AI governance, but emphasizes the need for the technology to respect "fundamental socialist values."

mundophone