OpenAI announced on Thursday that it has dismantled two covert influence operations originating from Russia and Iran. The campaigns utilized the company's artificial intelligence chatbot, ChatGPT, to generate fake journalist personas and operate a front organization designed to spread anti-Ukraine propaganda and stoke political division within both the United States and Israel regarding the war in Iran.
According to the artificial intelligence firm, the Iranian campaign relied on seven fabricated journalist profiles. These personas falsely claimed to be based in Western nations, including the United States, and were used to inject long-form articles into global media networks. OpenAI revealed that it tracked nearly 100 articles that were either published or syndicated under these fake bylines across roughly a dozen international online outlets. These platforms were characterized as small-to-medium websites primarily focused on Middle Eastern events, international affairs, and geopolitics.
One specific persona identified in the operation was "Ervin B. Hoskins," who was presented online as an "American freelance writer." This fake persona maintained social media profiles across multiple major technology platforms, including Meta's Instagram and Elon Musk's X. However, digital footprints exposed the deception. Transparency data from X indicated the account connected through a "West Asia android app," while Instagram's metadata placed the account's location directly in Iran. OpenAI confirmed that both social media accounts have since been suspended.
In a separate campaign, the Russian-backed operation co-opted unsuspecting individuals in Latin America to staff a front organization called the Social Research Center (SRC). This fake think-tank was used to execute covert actions to damage Ukraine's reputation across Latin America and sway local political landscapes. OpenAI noted that the evidence suggests the employees hired by the SRC in Latin America were entirely unaware that they were actually working for a Russian entity. Most of the original content published by the SRC was generated by this co-opted staff, representing what OpenAI described as "the most complex attempt to run a front identity that we've disrupted over the past two and a half years."
The Russian operators utilized ChatGPT to streamline their workflow, using the AI tool to write internal reports, draft performance assessments for the think-tank, and generate promotional content. The software was also used to draft social media posts and comments. Among their online activities, OpenAI highlighted efforts to exploit European political debates, such as generating reactions to Polish far-right MEP Grzegorz Braun's proposal to declare individuals exhibiting "anti-Polishness" as persona non grata. This specific activity was designed to amplify divisions within Poland regarding its ongoing support for Ukraine.
