From Foreign Policy · Mohamed Elgohari · · 5m
Iran's Lego videos are serious propaganda
A small Iranian team is using AI-generated animations to spread influence faster and cheaper than ever before, and Western officials are finally paying attention.
In brief
A small Iranian team is using AI-generated animations to spread influence faster and cheaper than ever before, and Western officials are finally paying attention. Generative AI has lowered the barrier to creating effective state propaganda, allowing tiny teams to reach billions with content that erodes trust rather…
Rubio dismisses Iran's Lego clips
In July, U.S. Secretary of State Marco Rubio called Iran's AI-generated Lego animation videos "stupid" while discussing negotiations. He wasn't even asked about them, revealing how much the campaign had penetrated American discourse.
“"stupid Lego videos they put out"”
Viral Lego Trump gets millions of views

In March, a video showed a Lego-figurine Donald Trump waking in the Oval Office as Iranian missiles strike a U.S. warship. It was viewed millions of times within days, produced by a small Iranian group called Akhbar Enfejari using consumer-grade hardware.
Small team, massive output
Akhbar Enfejari, operating from inside Iran with roughly 10 people, released over 100 videos across X, YouTube, Instagram, Telegram, and Spotify. YouTube suspended their channel in April, but the content kept spreading.
AI changes propaganda's economics
Generative AI eliminates the need for separate teams of writers, translators, designers, and animators. Polished content can now be created in hours by people without sophisticated technical skills.
One team, two audiences

Akhbar Enfejari produces English-language satire for Western audiences resistant to Iranian messaging, while simultaneously creating Persian-language videos reinforcing Iran's ideology for receptive regional audiences.
English videos don't need belief
The Lego clips aren't designed to persuade. They're watched, laughed at, and shared. They capture attention and feed on themselves, like when Rubio's dismissal prompted a new clip portraying him as "Little Marco."
Corrosion, not conversion
Research shows institutional trust erodes even when viewers reject the underlying claims. The goal is to damage confidence in institutions, not to convince people of specific falsehoods.
Lego format bypasses content moderation
Lego is globally recognized, aesthetically inoffensive, and carries no political charge. Violent scenes rendered in plastic bricks read as satire rather than incitement, slipping past systems designed to catch graphic content.
Videos fit existing internet tradition
Lego parody animations called brickfilms have circulated on YouTube for decades. Akhbar Enfejari's videos pass as entries in a recognizable tradition rather than state propaganda.
Persian videos use devotional sincerity
Running simultaneously, Persian-language videos feature Shiite mourning chants, invocations of Imam Ali, and references to the Iran-Iraq War. They adopt none of the irony of English tracks, maintaining legitimacy through religious and historical reverence.
Massive reach on minimal resources
Two pro-Iran networks generated over a billion views on X in the first month of the current war. Akhbar Enfejari approached that reach with a handful of people and consumer hardware while Iran's government maintained plausible deniability.
Trump adopts the same tactic
Trump posted AI-generated war videos depicting fictional strikes on Iran's Kharg Island. Vice President J.D. Vance said Trump was "sending a message to the Iranians," using the exact logic that makes Lego videos work: visibility over truth.
Model is highly replicable
The Akhbar Enfejari approach shows that a small, cheap, plausibly deniable team can now produce propaganda matching a state media apparatus's reach. The division of labor is simple enough for others to copy.
Deplatforming doesn't stop spread
YouTube suspended Akhbar Enfejari in April, but videos kept circulating through amplifier networks, state media accounts, and sympathizer reposts. The platform later lifted the ban, and the group now operates on multiple platforms.
QR codes create persistent pathways
Several videos embed QR codes directing viewers to affiliate channels, allowing reuploaded copies to retain connections back to the campaign even after the original channel is removed.
Labels don't reduce sharing
AI content labeling assumes audiences discount labeled material, but evidence shows labeling doesn't reliably reduce sharing once content has emotional traction. Multiple disposable versions from many nodes make labeling ineffective.
Track it like militant propaganda
Policymakers should treat this as a genre using tools built for tracking militant propaganda: narrative coding, audience targeting, and longitudinal comparison across releases applied systematically rather than fact-checked once.
Map the amplification networks
The real distribution engine is state media accounts, diplomatic handles, and sympathizer reposts. Effective response requires mapping these amplification networks rather than targeting individual channels.
Treat content as relationship-building
Countermessaging treats this as arguments to rebut, but the appeal is the audience relationship it builds and the ironic distance that makes sharing feel like commentary rather than endorsement.
Model will spread beyond Iran
Akhbar Enfejari has already been deplatformed and kept going. The model is already being copied by actors with no stake in Iran's war and won't stay confined to the conflict that produced it.
Here's the gist
Generative AI has lowered the barrier to creating effective state propaganda, allowing tiny teams to reach billions with content that erodes trust rather than persuade.


