OpenAI discloses and bans two AI-driven false-front influence operations on October 8
OpenAI disclosed two covert influence operations using ChatGPT, with the Russian Dark Clark operation rated Category 5 on Breakout Scale, the first such rating the company has issued.
OpenAI published a safety report on October 8 disclosing and banning two covert influence operations that used ChatGPT, codenamed Dark Clark from Russia and Bogus Bylines from Iran. The Russian operation is rated Category 5 on the Breakout Scale, the first Category 5 operation OpenAI has identified. Dark Clark targeted Latin America, used ChatGPT to draft internal reports, operated a false think tank called Social Research Center, and bypassed OpenAI's Russia access restrictions through VPN. The operation forged documents and audio from Peru's education authority, Ecuador's Ministry of Education and Bolivia's EPSAS water utility, with Argentina, Bolivia, Ecuador, Peru and Brazil as target countries. Bogus Bylines used seven fictitious journalist personas including Ervin B. Hoskins, Noah Lamington, Sophia Gonzalez, Michael Harrison, Ericka Feusier, Jenny Williams and Alice Johnson, publishing about 100 articles in Persian and English across roughly a dozen online outlets. The article publication portion was rated Category 4, while the social media comment portion was only Category 2. OpenAI noted that it has publicly disclosed 30 covert influence operations since early 2024, and that false content reaching real media outlets reaches wider audiences than content spread only on social platforms.