OpenAI Blocks Russian Influence Campaign Using Fake Israeli Think Tank

OpenAI has uncovered and blocked a covert Russian influence operation that used ChatGPT to promote a fictitious research institute posing as Israeli. The campaign employed VPNs to bypass service restrictions in Russia and directed the AI to generate posts and comments in English and German while concealing the operators' Russian identity. These materials were disseminated across social media platforms including X, Facebook, LinkedIn, Substack, and Telegram.

Central to the scheme was the International Burke Institute (IBI), a fake think tank claiming a physical address in Israel and presenting itself as a prestigious community of global intellectuals. OpenAI's investigation of 36 articles on the IBI website found that 34 were plagiarized from legitimate academic sources but attributed to false authors, alongside poorly machine-translated texts from Slavic languages. The site promoted a fabricated "sovereignty index" that portrayed Russia positively and systematically attacked Western countries opposing the war in Ukraine, including France, Germany, the US, and the EU.

Additionally, operators created profile images and logos for dozens of Telegram channels spreading disguised pro-Russian propaganda. A troubling element involved a small number of real Israelis who allegedly represented the IBI in submitting papers and attending professional conferences, though OpenAI could not determine their relationship with the Russian network or whether they acted in good faith.

Despite the sophisticated setup, OpenAI estimates the campaign's reach was limited, with most posts receiving only a few views and Telegram channels having only a few thousand followers. The company warns this incident marks a dangerous escalation in state actors leveraging AI tools to fabricate scientific authority, mask propaganda sources, and undermine information sovereignty online.