OpenAI Says China Used ChatGPT in Covert Propaganda Campaigns Across Scores of Platforms
Updated
Updated · The Atlantic · Jul 7
OpenAI Says China Used ChatGPT in Covert Propaganda Campaigns Across Scores of Platforms
3 articles · Updated · The Atlantic · Jul 7
Summary
OpenAI said Chinese actors used ChatGPT to generate and spread English-language content aimed at turning Americans against AI data centers, then banned the accounts after the campaign gained little traction.
The company said the operators were likely tied to a private tech firm working for provincial officials and used the chatbot to create polarizing comics and comments that could slow U.S. AI infrastructure build-out.
In February, OpenAI also linked a Chinese user tied to law enforcement to an effort to discredit Japanese Prime Minister Sanae Takaichi, describing a sustained operation with at least hundreds of staff and thousands of fake accounts.
Researchers and OpenAI say AI is making Beijing’s influence operations and censorship faster, cheaper and more precise, while Chinese state media in training data can skew chatbot answers toward pro-China narratives.
A Nature study found DeepSeek answered political queries more favorably to China than ChatGPT 99% of the time, underscoring how wider adoption of Chinese AI tools could amplify Beijing’s messaging abroad.
In mid-2026, OpenAI released a report revealing two covert influence campaigns that used ChatGPT to target U.S. public debates—one about AI data centers and electricity prices, and another about U.S. tariffs and technological competition. OpenAI traced both operations to actors linked with China, who used various tactics to shape online narratives. Despite these efforts, the campaigns achieved minimal engagement and had no significant impact on public opinion. As a result, OpenAI banned the involved ChatGPT accounts, highlighting the ongoing concerns about foreign influence in U.S. technology discussions.