OpenAI has banned multiple ChatGPT accounts linked to a China-origin influence operation, dubbed "Uncle Spam," following a tip from Meta. This operation generated polarized social media content, supporting both sides of divisive U.S. political topics, including text and AI-generated profile images on platforms like X and Bluesky.

The network utilized OpenAI's models to create social media comments, particularly regarding tariffs, and even sought advice on optimal posting times for maximum engagement. They also used image generation for profile logos and to create personas, notably of U.S. veterans critical of the current administration. Furthermore, the accounts attempted to extract personal data from social platforms and sought guidance on analyzing profile information for categorization and inference.

The content produced was polarizing, often within the same operational sessions, suggesting an intent to exploit existing divisions rather than promote a specific ideology. OpenAI assessed the impact as Category 2, with minimal authentic engagement despite some accounts having thousands of followers. The operation's focus on American messaging and the volume of generated content led to the "Uncle Spam" designation.