TL;DR
Iran, China, and Israeli private firms deployed AI agents built on freely available Chinese models to run influence campaigns on US social media, marking the first known near-fully automated foreign interference operations.
What happened
- Hundreds of AI agents, powered by publicly available Chinese AI models, were deployed to create and operate fake social media accounts across Instagram, Facebook, X, and TikTok.
- The agents autonomously generated posts, coordinated messaging, and engaged with political content, requiring minimal human oversight throughout the process.
- The Iranian campaign targeted US audiences, with AI-generated accounts posing as ordinary Americans in major cities, posting memes, tagging journalists and politicians, and promoting anti-Republican content.
- Nearly 80,000 people followed the Iranian AI-generated accounts, which were active during the first half of 2025.
- A separate campaign linked to Tel Aviv-based IntelEye used AI agents to control thousands of accounts engaging with discussions around Israeli elections; the company called it a defensive research experiment.
Why it matters
- US officials and researchers described these as among the first known cases where AI agents handled nearly the entire influence operation pipeline, from account creation to content posting to audience engagement.
- The campaigns were crude but consequential: they prove that publicly available AI tools can now automate what previously required large, expensive human troll farms.
- State-backed actors in Iran and China were traced as responsible, raising the threat level from nuisance to strategic national security concern.
- The use of freely available Chinese AI models means the barrier to entry for this kind of operation is now close to zero for any motivated actor, state or otherwise.
- Social platforms face a novel detection challenge: AI-generated accounts that post, interact, and coordinate at scale look far more organic than older bot networks.
What to watch next
- Whether Meta, X, TikTok, and Instagram disclose specific takedown numbers and the detection methods they used, which would signal how prepared platforms are for the next wave.
- How US intelligence and the Biden or successor administration respond with policy, sanctions, or platform mandates targeting AI-enabled foreign influence operations.
- Whether Chinese AI model providers face pressure to add export controls or usage restrictions, given their models were the engine of these campaigns.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.