OpenAI disabled a network of ChatGPT accounts linked to a Russian covert influence operation. These users employed VPNs to bypass regional access restrictions. The group instructed the AI to generate multilingual social media comments while concealing any Russian linguistic markers.
The operation promoted the International Burke Institute (IBI), a fraudulent Israeli-based think tank. This network spread pro-Kremlin narratives across X, Facebook, LinkedIn, and Telegram. Content typically praised Russia while criticizing Ukraine, the United States, and the European Union.
OpenAI reported that the campaign reached a relatively small audience despite its elaborate organizational structure.