From the source
OpenAI banned two covert influence operations using ChatGPT, one from Russia (assessed at Category 5) and one from Iran (Category 4), which used fake front entities to launder geopolitical messaging into target audiences.
- The Russian operation, dubbed 'Dark Clark,' targeted Latin America with false stories and a fake research platform, while the Iranian campaign used fake journalist personas to pitch articles to online outlets.
- OpenAI has disrupted 30 such operations over the past two and a half years.