AI-powered influence campaigns manipulate social media in US and Israel
State-backed groups in China and Iran, along with Israeli private firms, allegedly used openly available Chinese AI models to create autonomous agents that operated thousands of fake social media accounts spreading false political information. US intelligence agencies and security researchers traced the campaigns to state-backed actors and private companies, with nearly 80,000 people following Iranian AI-generated accounts targeting US audiences.
What happened
According to a New York Times investigation, state-backed groups in China and Iran, along with private firms in Israel, used openly available Chinese AI models to create autonomous agents that operated thousands of fake social media accounts.
The AI agents autonomously opened fake accounts on Instagram, Facebook, X/Twitter, and TikTok, then populated them with false information relating to politics and current events. The agents bypassed platform safeguards designed to prevent account manipulation.
US officials said this marked one of the first times Chinese-produced AI agents were used to handle every stage of an online influence campaign, from creating fake accounts to coordinating messaging and framing online.
Reported impact
The Iranian campaign targeted US audiences with AI-generated accounts posing as ordinary Americans. Nearly 80,000 people followed the AI-generated accounts, which posted popular memes, tagged journalists and politicians, and promoted views critical of the Republican Party.
An Israeli campaign linked to Tel Aviv-based IntelEye involved AI agents controlling thousands of social media accounts and engaging with discussions surrounding Israel's upcoming elections. The company said the activity was a defensive research experiment designed to test how AI models could be misused.
Index analysis
The campaigns were described as relatively crude but significant because they demonstrated how publicly available AI systems can automate operations that previously required substantial human involvement. The open-source AI models were manipulated by developers to allow agents to evade safeguards designed to verify users and prevent online discourse manipulation.
A fellow at the Georgetown Center for Security and Emerging Technology noted that 'Agent-led campaigns' will likely become more common as the technology spreads, as agents can be sophisticated and blend in to look more like human users.
Governance controls
US intelligence agencies, technology companies and security researchers investigated the activity and traced the Chinese and Iranian operations to state-backed actors.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.