OpenAI disrupts covert influence campaigns tied to China, North Korea, Russia and Iran
OpenAI reported that it disrupted covert influence and cyber operations by China, North Korea, Russia and Iran using ChatGPT and Sora for propaganda, social engineering, phishing and surveillance. The company banned the associated accounts and stated there is little evidence the AI use increased operational reach compared to traditional methods.
What happened
OpenAI released a threat report on June 5, 2025 stating it had disrupted multiple covert influence and cyber operations tied to foreign actors including China, North Korea, Russia and Iran, and banned associated accounts.
Chinese actors used ChatGPT to create content for social media platforms including TikTok, Reddit, X and Facebook, generating posts criticizing U.S. foreign policy and tariffs, impersonating both sides of American culture war debates, and creating fake sock puppet accounts. Operationally named campaigns included 'Sneer Review', which mass-produced short political comments in multiple languages, and attacks on a Taiwanese video game called 'Reversed Front' that contains anti-CCP themes.
Chinese operators also used ChatGPT to generate phishing emails, translate propaganda into Spanish for Latin American outlets, create fake biographies for fake social media accounts posed as journalists, and produce internal-style documents on self-discipline for Chinese public security organizations.
Reported impact
OpenAI representatives told reporters there is little evidence that the use of AI resulted in greater reach or impact than traditional methods.
Governance controls
OpenAI shut down a number of covert influence and cyber operations and banned accounts tied to malicious use of its AI models, halting four operations likely tied to the Chinese communist regime.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.