OpenAI pulled the plug on two AI-powered propaganda operations that built fake journalists and a phony think tank to peddle geopolitical talking points.
The company said it identified and disrupted two separate influence operations that relied on its AI tools. One operation invented journalist personas - complete with bylines - to produce and distribute content. The other set up a false-front think tank, using AI to generate reports and commentary designed to look like independent analysis. Both were aimed at spreading geopolitical messaging rather than selling a product or running a financial scam.
This is the latest entry in OpenAI's running tally of busted influence operations, and it underlines a quieter truth about generative AI: the scariest misuse case isn't a chatbot giving bad advice, it's AI manufacturing the appearance of independent voices. A fake think tank with convincing output is cheap to spin up and hard for readers to tell from the real thing.
Whether OpenAI is catching a meaningful share of this activity or just the sloppy operators who get caught is the harder question - and one the company's self-reported takedowns can't really answer.