Chinese government-linked actors used AI tools including generative AI for imagery and AI-generated news anchors to conduct disinformation campaigns targeting U.S., Taiwan, and other foreign elections through fake social media accounts and websites.
Chinese government-linked online actors, primarily through the Spamouflage operation (also known as Storm-1376 and Dragonbridge), increasingly leveraged artificial intelligence tools to target voters in the U.S., Taiwan and elsewhere with disinformation campaigns. The campaigns used fake social media accounts on platforms like X, some over a decade old, to post about divisive American topics including drug use, immigration, and racial tensions, and asked followers to share opinions about presidential candidates. The operations used relatively rudimentary generative AI for imagery and more sophisticated AI tools including AI-generated news anchors created using ByteDance tools, fake audio clips of political candidates, and AI-generated images of disasters like the Maui wildfires. Microsoft researchers identified this as among the first direct ties between generative AI tools and covert state-sponsored online influence operations against foreign voters. The campaigns spread false conspiracy theories about events like train derailments and wildfires, alleging U.S. government responsibility. While the posts failed to achieve significant traction according to Microsoft, they demonstrated China's experimentation with AI tools for election influence. The operations targeted multiple countries but showed particular intensity around Taiwan's January 2024 presidential election, where AI-generated content volume expanded significantly in recent months.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to conduct large-scale disinformation campaigns, malicious surveillance, or targeted and sophisticated automated censorship and propaganda, with the aim of manipulating political processes, public opinion, and behavior.
AI system
Due to a decision or action made by an AI system
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed