OpenAI has banned ChatGPT accounts suspected of being used for a covert influence operation originating from China. The activity involved bulk-generating social media posts and drafting internal performance reviews. Prompts were primarily in Chinese, focusing on political and geopolitical topics relevant to China. One user claimed to work for the Chinese Propaganda Department, though this could not be independently verified.

The banned accounts engaged in two main workstreams. The most prolific was generating short social media comments in English, Chinese, and Urdu, posted on platforms like TikTok and X to create a false impression of organic engagement. This often involved a main account posting followed by numerous reply comments from other accounts. These accounts used varied screen names, sometimes unrelated to the content's language, and cartoon profile images on X.

The second workstream involved using ChatGPT to generate internal policy or performance review documents. The social media content aligned with China's geostrategic interests, targeting topics such as Taiwan-centric games, criticism of Chinese investments in Balochistan, and commentary on the closure of USAID. The operation appeared to be in its early stages when detected, with limited organic engagement on most platforms, though some TikTok videos and X tweets showed higher view counts, with the caveat that much of the engagement was likely inauthentic.