OpenAI says it shuts down two influence operations, one tied to Russia and one to Iran, that used the company’s chatbot and models to support covert political messaging. In a safety report published Thursday, OpenAI describes the networks as using “false front” entities intended to give cover to coordinated propaganda efforts.
According to multiple outlets, both operations relied on creating fake journalist personas and presenting content through a purported think tank. France 24 and others report the Iranian effort involved several profiles posing as Western journalists, while the Hill and CyberScoop say OpenAI identifies the Russian operation as among the widest-reaching schemes it has recorded in these disruptions. OpenAI also frames the incident as an example of how state-linked actors can exploit readily available AI tools at scale.
Outlets differ mainly in emphasis and detail: some highlight the anti-Ukraine focus of the messaging, while others stress OpenAI’s rating of the campaigns’ severity. CyberScoop notes OpenAI’s assessment marks the first time it reported what it considers high-impact influence operations, while other coverage focuses on the mechanics of using AI-generated “media” identities to disseminate narratives.
