AI threatening Party rule in China? Here's what Xi Jinping's govt is doing

China's worried that Chatbot's ability to generate independent responses could lead people to question Communist Party rule and official narratives

Artificial intelligence (AI) is rapidly gaining influence worldwide, including in China. Even as Beijing positions itself as a global leader in AI, it is growing increasingly concerned about the technology’s impact on the country’s political system, according to a report by The Wall Street Journal.

 

The report said chatbots pose a particular challenge for the Chinese government. Their ability to generate independent responses could lead people to question Communist Party rule and official narratives.

 

Recognising these risks, the Chinese government has moved quickly to tighten control over AI use. Earlier this year, Chinese President Xi Jinping warned that AI brings “unprecedented risks,” comparing unsafe AI to "driving on a highway without brakes."

 

 

Let's take a look at how China is trying to keep AI under control.

How is the Chinese government controlling AI?

China is tightly regulating both how AI systems are trained and how their outputs are shared with the public.

 

According to The Wall Street Journal, under new national AI standards implemented last month, companies must closely screen their training data. Human reviewers are required to examine thousands of data samples, and at least 96 per cent of the content must be considered safe before it can be used.

 

The rules list 31 categories of risk, with the most serious being content that encourages the overthrow of state power or the socialist system.

 

In addition to data controls, strict regulations apply to chatbots and generative AI tools. The latest rules require chatbots to be trained only on data that has been filtered for politically sensitive material. Before they are released, AI models must pass an official ideological review.

 

All AI-generated text, images and videos must also be clearly labelled and traceable, making it easier for authorities to identify and punish users who share content deemed undesirable.

 

During a recent three-month enforcement drive, Chinese officials said they removed about 960,000 AI-generated items that were classified as illegal or harmful, the report said.

How will this affect Chinese AI firms?

Despite strict controls, Chinese AI models have performed well in international rankings, including in areas such as computer coding. However, they often avoid or censor responses related to sensitive issues such as the Tiananmen Square massacre, human rights, and Taiwan.

 

DeepSeek, despite its rapid rise earlier this year, has faced criticism for its handling of political questions. Users have reported that its Chinese-language version often avoids political topics or closely follows official government narratives.

 

Additionally, most major American AI models remain unavailable in China. Experts say it could become more difficult for companies such as DeepSeek to keep pace with US rivals as AI systems become more advanced and complex.