AI Agents Now Automate Foreign Influence Campaigns on Social Media

AI agents take foreign influence operations to new scale

Foreign actors are now using AI agents—software systems capable of completing multiple tasks with minimal human intervention—to run influence campaigns across social media platforms. According to reporting from The New York Times, Iran and China, along with private Israeli companies, have deployed these systems to create fake accounts, generate political content, and coordinate messaging across Facebook, Instagram, X, and TikTok.

The shift represents a significant evolution in how influence operations work. While AI-generated content and fake personas have existed for years, the newer campaigns use agentic AI that can execute entire workflows automatically. In documented Iranian operations, hundreds of AI agents created networks of fictitious accounts posing as ordinary Americans in major U.S. cities. These accounts posted memes and political commentary, tagged journalists and politicians, and promoted messages critical of the Republican Party, accumulating nearly 80,000 followers during the first half of 2026.

Why it matters

Automation makes influence campaigns cheaper to run and easier to scale, allowing foreign actors to create the illusion that fringe viewpoints enjoy widespread support. Rather than producing isolated propaganda pieces, adversaries can now deploy entire networks of coordinated accounts designed to exploit existing political divisions. U.S. officials including the FBI and CISA have warned that these operations aim to sow discord and increase polarization rather than simply advance one political agenda.

Reach versus persuasion

The documented campaigns have achieved significant visibility. The Institute for Strategic Dialogue tracked more than one billion views generated by two coordinated pro-Iran networks on X during the first month of the Iran war. These networks circulated false, misleading, and AI-generated material that spread through recommendation algorithms, reposting by larger accounts, and coordinated amplification.

However, researchers emphasize an important distinction: views and follower counts measure distribution, not persuasion. A post receiving millions of views indicates how widely it appeared in feeds, but does not establish how many people believed the content or changed their political opinions as a result.

What's actually new

The technology itself isn't entirely novel—AI has been used for years to generate articles, images, videos, and social media posts for influence operations. The development that concerns researchers is the degree of automation. Agentic AI can now carry out complex, multi-step operations with less human oversight than previous systems required.

Anthropic's September 2026 threat-intelligence report documented similar trends, noting the expanding use of AI in malicious influence operations and other forms of online abuse.

For social media users, the implication is straightforward: a post that appears to come from an ordinary person may actually originate from an automated system designed to manipulate political discourse. The documented campaigns show foreign actors are actively experimenting with ways to automate influence at scale, though measuring the actual impact on American political beliefs requires evidence beyond metrics like follower counts or view tallies.

The details were first reported by The New York Times.