OpenAI shut down China linked accounts that used ChatGPT to generate content aimed at shaping US political and technology debates.

The activity included posts, slogans, comments, and cartoons focused on President Donald Trump’s tariffs, US technology policy, AI infrastructure, and data centers. OpenAI said the campaigns dated from late 2025 into early 2026 and appeared to have little or no effect.

One group used ChatGPT to generate social media content criticizing Trump’s trade and technology agenda. Some of that content was later posted to X, while the same users also generated Chinese language comments and material in other languages.

OpenAI also traced another group of users to a Chinese technology company that had performed government work. That operation focused on US debates over AI and data centers, portraying the industry as costly, energy intensive, and harmful to ordinary citizens.

The company has internally labeled one China linked influence effort as Uncle Spam, part of a broader pattern of foreign actors testing generative AI for propaganda, fake personas, cyber activity, and social media manipulation.

The low engagement matters. OpenAI rated similar influence operations as low impact when content appeared across platforms but showed little evidence of being picked up or widely shared by real users.