Iran, China, and groups in Israel have recently employed artificial intelligence in unprecedented ways to conduct social media influence campaigns, escalating concerns among U.S. officials and security researchers regarding the rapid advancement of this technology.
These state and non-state actors combined AI tactics to significantly increase the speed and scale of online campaigns, operating with minimal human intervention, according to U.S. officials and security researchers familiar with these operations.
The process involved utilizing powerful, publicly accessible Chinese AI models. These models were then used to create "agents"—autonomous bots capable of performing tasks and operating software independently. Once deployed, hundreds of these AI agents autonomously established networks of fake accounts across platforms like Instagram, Facebook, X, and TikTok. These agents then rapidly populated the fake accounts with disinformation concerning political and current events to manipulate public opinion and exacerbate societal divisions.
U.S. officials, who spoke on condition of anonymity due to the sensitive nature of the information, indicated that this marks one of the earliest instances where AI agents, powered by Chinese models, managed every phase of an influence campaign, from account creation to coordinated messaging. They also noted that safeguards typically designed to prevent AI models from generating fake accounts or manipulating online discussions were deactivated by the operators of these systems.
While described as rudimentary, these "agentic" campaigns attracted the attention of U.S. intelligence agencies, major technology companies, and security experts. Investigations into these efforts revealed that the Chinese and Iranian campaigns were state-sponsored, while two Israeli campaigns were linked to private companies.
