An account, likely originating in the United States, has been banned by OpenAI for creating a fake ChatGPT error message designed to appear as if it detected "Russian troll" activity. The hoax post, which went viral on X (formerly Twitter), claimed to show an error message from a Russian-speaking user running out of credits while trying to generate pro-Trump content.

OpenAI's investigation revealed that while the account had previously used their models for adversarial comments on various topics, the specific error message was not generated by their AI. Instead, it appeared to be manually created and contained invalid code. The account's behavior involved generating argumentative replies to other users, often on trivial subjects, with a common theme of being contrarian.

This incident is notable because it involved non-AI activity being used to deceive people about the use of AI models, rather than the other way around. The hoax quickly gained significant traction, spreading across multiple social media platforms and attracting media attention, highlighting how rapidly a compelling hoax can be amplified online. The situation also underscores a shift in public perception, with a greater skepticism towards alleged Russian influence operations.