OpenAI has identified and banned several ChatGPT accounts linked to a covert online influence operation originating from Russia.
This operation aimed to disseminate misinformation through AI-generated social media posts, which were part of a broader campaign that included a website featuring misattributed academic work and a 'sovereignty index' that portrayed Russia positively while criticizing Western nations.
Although OpenAI does not permit access to its models in Russia, the use of VPNs allowed these accounts to operate undetected. The campaign's immediate impact may have been limited, but OpenAI emphasized the importance of the infrastructure established by these operators, which could potentially be scaled for future misinformation efforts.
This incident underscores the risks associated with AI technologies being exploited for influence and the challenges in monitoring and regulating such activities