Iran, China Deploy First Autonomous AI Influence Campaigns Across Social Media Platforms

State-Backed Actors Deploy Autonomous AI Agents

Iran, China and Israeli groups have created first-of-their-kind AI influence campaigns on social media, marking a significant escalation in how nations manipulate online discourse1

1

. U.S. officials and cybersecurity researchers confirmed that these campaigns combined AI-driven disinformation tactics to amplify the speed and scale of operations beyond what humans could achieve, with minimal human intervention required1

1

.

The actors turned to Chinese open-source AI models, which are freely available systems that can be downloaded and modified without restrictions. They then used these models to create autonomous AI agents—bots capable of operating software and completing tasks independently1

1

. Once deployed, hundreds of these AI agents autonomously opened networks of fake social media accounts on platforms including Instagram, Facebook, X and TikTok2

2

.

AI-Enabled Influence Operations Run End-to-End

This represents one of the first times AI agents backed by Chinese models handled every stage of influence campaigns, from creating fake accounts to coordinating messaging online, according to U.S. officials not authorized to speak publicly on sensitive issues1

1

. The agents rapidly seeded fake social media accounts with false posts about politics and current events designed to sway opinions and inflame issues1

1

.

In most cases, safeguards that would normally prevent an AI model from creating a fake account or manipulating online discussions were deliberately disabled by those operating the systems1

1

. Unlike proprietary models from companies like OpenAI and Anthropic, open-source AI models can run on private hardware, meaning no company can intervene to shut the systems down1

1

.

U.S. Intelligence and Tech Companies Respond

The campaigns, while crude in execution, caught the attention of U.S. intelligence agencies, major tech companies and security experts who worked to detect the actors behind the efforts and their methods1

1

. Investigations traced the Chinese and Iranian campaigns to state-backed actors, while two Israeli campaigns were linked to private companies1

1

.

Meta identified what it called a "technically significant development" of bad actors experimenting with AI and confirmed it had detected Iranian and Israeli campaigns using automated content production1

1

. The company stated that AI is now "embedded within automated systems that can produce, adapt, and distribute content with minimal human intervention"2

2

. Meta told The New York Times that the vast majority of AI-powered agents set up on Facebook and Instagram by IntelEye, other companies, and state-backed actors had been removed by its automated detection and verification systems2

2

.

Israeli Company Claims Defensive Research

IntelEye, a Tel Aviv-based company founded by former NSO Group employees, operated roughly 1,000 AI-powered fake accounts on social media around the upcoming Israeli elections2

2

. The company used DeepSeek, a Chinese open-source AI model, to create agents that published content both supporting and opposing Prime Minister Benjamin Netanyahu2

2

.

IntelEye disputed claims that its work was intended to influence online discussions. Co-founder Maor Sellek stated, "IntelEye does not operate influence campaigns. Our work with DeepSeek was conducted for defensive research and testing, to understand how models with insufficient safeguards could be misused and to improve detection of that activity"2

2

. Sellek said IntelEye bought 10,000 social media accounts and handed over control to autonomous AI agents, though only about 1,000 were operational2

2

.

Concerns Mount Ahead of U.S. Midterm Elections

The campaigns have raised fears over how publicly available AI models can be weaponized to carry out AI-enabled influence operations essentially on their own. Some U.S. officials expressed concern that similar tactics could target U.S. voters ahead of the midterm elections in November1

1

. Russia, Iran and China have all targeted American voters with online influence campaigns during past election cycles, each using at least rudimentary AI1

1

.

Kyle Crichton, a fellow at the Georgetown Center for Security and Emerging Technology who researches AI and cybersecurity, told The New York Times that "agent-led campaigns will probably become more common as the technology continues to improve"1

1

. He added, "Agents can be fairly sophisticated, and so in addition to just scale, they have an ability to blend in and look more like a human user"2

2

.

Kyle Chan, a researcher at Brookings, noted that agents made by companies like OpenAI have been misused before, including by Iran as early as August 2024, but open AI models allow for a new level of autonomy1

1

. Watch for increased regulatory scrutiny around open-source AI models and their potential for misuse in political misinformation campaigns as AI safety debates intensify.