OpenAI pulled the plug on accounts caught using its tools to fake support around Ghana's presidential vote.
The company banned a cluster of accounts that used its AI to generate articles, social media posts, and artificial engagement tied to Ghana's 2024 presidential election. The activity was designed to look like organic commentary rather than a coordinated campaign. OpenAI did not name the operators or say which candidates the content favored. The takedown is one entry in a running list of influence operations the company says it disrupts as it finds them.
Elections in smaller democracies draw far less scrutiny from platform trust and safety teams than US or EU races, which makes them a cheaper testing ground for AI-generated propaganda. Banning accounts on one company's tools only closes one door - the same operators can move to another model or hire humans to keep writing. That's the actual limit of this kind of enforcement.
It's the same pattern OpenAI has flagged elsewhere: generate volume, fake the appearance of grassroots support, and hope nobody checks the account history.