OpenAI has moved to disrupt several accounts associated with the illicit use of its ChatGPT chatbot for surveillance and influence operations, the most notable of which is a suspected China-based social media listening tool that leverages the firm's AI models to obtain information on anti-China demonstrations in the West, according to The Hacker News.
Also dismantled by OpenAI were a likely Chinese network associated with the Spamouflage attack campaign producing English-language social media content and Spanish-language long-form articles attacking the U.S., a North Korean fake IT jobs scheme, and numerous accounts used by North Korean state-backed threat operations Kimsuky and BlueNoroff in intelligence efforts.
OpenAI also banned accounts involved in a romance-baiting scam, Iranian and Ghanian influence operations, and a task scam believed to have been based in Cambodia.
Such a development comes after dozens of Chinese-, Russian-, North Korean-, and Iranian-linked hacking operations were reported by Google's Threat Intelligence Group to have exploited the Gemini AI chatbot in their respective attacks.
