Technology
OpenAI bans Russian ChatGPT accounts used in covert online influence campaign
The operation used AI to build fraudulent academic credentials, bypass regional restrictions, and post content across multiple social platforms.
The short version
- OpenAI terminated a network of Russian ChatGPT accounts that utilized VPNs to bypass regional access restrictions and generate social media content.
- The campaign created a website called the International Burke Institute with copied academic material and a 'sovereignty index' designed to praise Russia and criticize Western nations.
- OpenAI reported that while the operation's immediate audience reach appeared limited, it demonstrated how AI can be used to manufacture authority for covert campaigns.
Key facts
- OpenAI banned a network of ChatGPT accounts originating in Russia after an investigation into AI-generated social media posts uncovered a broader covert influence operation.[CNBC]
- The campaign operators used virtual private networks to circumvent OpenAI's geographical access limits in Russia and instructed the model to omit Russian linguistic markers.[CNBC]
- Generated content was posted across several platforms, including Substack, Telegram, X, Facebook, and LinkedIn.[CNBC]
- The operation established a website for an entity called the International Burke Institute featuring plagiarized and misattributed academic articles along with a proprietary sovereignty index.[CNBC]
- In February, OpenAI previously shut down ChatGPT accounts linked to Rybar, a pro-Russia media company.[CNBC]
What remains uncertain
- The full identity of the individuals or organization operating the campaign remains unverified beyond their origin in Russia.[CNBC]