Chinese Government Uses AI Tools for Propaganda Campaigns

Chinese operatives used an American AI chatbot to create comics and social-media posts that highlighted the energy demands of data centers, according to OpenAI. The company stated that the accounts appeared linked to a private technology firm working for provincial government officials. OpenAI banned the accounts after the posts appeared.

In a separate case, a user connected to Chinese law enforcement asked the same chatbot to design a campaign criticizing Japanese Prime Minister Sanae Takaichi on immigration and living costs. The chatbot refused, but OpenAI described the request as part of a larger effort involving hundreds of staff and thousands of fake accounts.

The operation targeted critics of China by creating false social-media profiles, flooding platforms with pro-government posts, and spreading inaccurate information about dissidents.

Researchers at the Australian Strategic Policy Institute reviewed corporate records and found that AI systems allow Chinese authorities to scan large volumes of media and remove restricted content more quickly. A study published in Nature reported that state-run Chinese news sources influence chatbot responses to questions about China when the queries are in Chinese.

The same study showed that a leading Chinese chatbot produced answers more favorable to China than the American model in 99 percent of tested political queries.