OpenAI Bans Russian and Iranian Accounts for AI-Driven Influence Operations
OpenAI said it removed Russian and Iranian accounts that used ChatGPT to assist influence campaigns across several countries. The Russian operations targeted Latin America, aiming to weaken support for Ukraine and affect local politics, and one campaign impersonated a Peruvian education authority. OpenAI reported that efforts to place false narratives in established media reached broader audiences than fake social media accounts.
OpenAI said its ban covered Russian and Iranian accounts that leveraged ChatGPT for influence work in several nations. Russian activity centered on Latin America, seeking to erode backing for Ukraine and shape local politics. Operatives largely used the chatbot to draft internal updates for superiors, according to OpenAI.
One Russian effort impersonated a Peruvian education office and persuaded schools to hold Ukraine-related events featuring Stepan Bandera, a WWII-era Ukrainian nationalist both admired and condemned. A Peruvian outlet reported an event occurred; Polish coverage prompted a far-right European Parliament member to propose entry restrictions for Bandera supporters. OpenAI also noted a fake Latin American think tank and Iranian Persian/English content.
The disclosure may heighten public scrutiny of generative AI's role in political persuasion, affecting platform users, journalists, and policymakers who must judge online claims. Latin American and diaspora audiences could encounter narratives tied to Ukraine's history, potentially deepening existing tensions. Newsrooms may face pressure to verify sources and AI-generated material, while regulators could consider stricter transparency rules. The reported reach of media placements suggests influence efforts may shape public debate beyond social platforms.