OpenAI has banned a network of ChatGPT accounts originating from Rwanda that generated partisan tweets in the lead-up to the country's elections. The activity was linked to individuals in Rwanda, with initial reporting from Clemson University researchers.

The network produced batches of short comments, often including hashtags, posted on X. Some accounts exhibited high-volume tweeting, with hundreds of posts per hour, and frequently duplicated the same message across multiple accounts. After initial disruptions, newly created accounts with similar content were also banned.

The operation focused on promoting the Rwandan Patriotic Front party, using hashtags like #RPFOnTop, #PKNiWowe, and #ToraKagame24. Posts were primarily in English, but also in French and Kinyarwanda. The repetitive nature of the content, due to a finite pool of synonyms and structures for specific themes, is a characteristic that can expose such operations.

While the network posted thousands of tweets, engagement was minimal, with single-digit replies, likes, or shares. This high-scale, low-engagement pattern is typical of comment-spamming networks. One objective may have been to make promoted hashtags trend on X in Rwanda, a tactic observed in pre-AI operations. The promoted hashtags did appear in the top ten trends at times, though not solely due to this network's activity.