The War in Plastic - Decipon Insights

The Video

On March 10, 2026, Iranian state media broadcast a short animated film. A Lego version of Donald Trump sits at a desk, panic-reading a folder labeled “Llrey Epstein File” — the AI-generated text slightly garbled, as it often is. A Lego Netanyahu cackles beside him. A figure representing Satan urges them forward. Trump launches a missile. It hits a girls’ school.

The animation is crude by Hollywood standards. The production value is roughly equivalent to a competent YouTube channel. The music is layered: trap beats underneath, swelling cinematic strings over the action sequences. The whole thing runs under three minutes.

“Narrative of Victory” became one of the most-viewed pieces of war propaganda of the century so far. In the first weeks of the conflict, AI-generated Lego videos from the same production network accumulated 145 million views. They spread on X, TikTok, Instagram, and YouTube — shared by pro-Iranian accounts, by Western anti-war accounts, and by ordinary users who simply found them striking.

The Iranian government had found the format it was looking for.

The War Context

The Iran war began February 28, 2026, when US and Israeli forces struck Iranian nuclear and missile facilities. The information environment in the days following was immediately saturated with synthetic media: deepfake videos of missiles hitting Tel Aviv, AI-generated footage of attacks on Bahraini high-rises, fabricated casualty figures. Most of those fakes were designed to look real and failed on contact with fact-checkers.

The Lego videos were not designed to look real. They were designed to look like something else entirely.

The Production Network

The videos carry the branding of a group called Explosive News Team — a self-described student media collective operating across X, TikTok, and until April, YouTube and Instagram. In email correspondence with journalists, the group claimed to be independent. When pressed in a BBC interview, a representative confirmed that the Iranian government was a “customer” — that official commissions had been placed for several projects.

Researchers and media outlets also linked the operation to the Revayat-e Fath Institute, a media body connected to the Islamic Revolutionary Guard Corps. The institute’s name translates, roughly, as “The Narration of Victory” — the same title as the March 10 video that went viral.

This is a familiar attribution structure: an institutional sponsor with state security connections, a nominally independent production front that claims grassroots origins, and a “customer” relationship that provides plausible deniability without severing the command link.

The videos themselves — The New Yorker called them “inescapable artifacts” of the war — were produced at the pace that only AI generation enables. A dozen-plus videos followed the initial hit, each iterating the format. Some added trap music prominently. Some leaned into the Minab school strike, where US munitions likely killed more than 100 civilians, most of them girls: small plastic shoes, a backpack in the rubble, rendered in Lego geometry. YouTube and Instagram terminated the accounts in April 2026. By then, the view count was already built.

Why Lego

The Lego format is not incidental to the propaganda’s effectiveness. It is the mechanism.

Lego is not a neutral aesthetic. It is one of the most culturally loaded visual languages in the world — associated globally with play, creativity, childhood, and The Lego Movie franchise. When you see Lego figures, your brain applies a set of pre-loaded associations: this is playful, this is for everyone, this is not serious in the way that matters.

This is the dual-process problem. Analytical cognition — the mode that evaluates claims, checks sources, identifies logical gaps — is slow, effortful, and easily bypassed by the right emotional or aesthetic triggers. Play mode activates a different orientation: open, receiving, pattern-enjoying rather than pattern-interrogating.

The format deploys this mechanism deliberately. A piece of content that opens with Lego figures is processed in play mode before the political payload arrives. The trap beats prime the body for emotional engagement. The cinematic orchestral swells attach emotional weight to the images. By the time Trump is launching missiles at children, the viewer has already been moved through three layers of cognitive conditioning — aesthetic, rhythmic, and emotional — without being asked to evaluate a single factual claim.

A researcher who analyzed the content for Time magazine described it precisely: “Legos tell your brain to play, the music tells your body to feel, and by the time the political payload arrives, you’re already in a mode where critical scrutiny doesn’t activate.”

This is not a trick that works only on unsophisticated audiences. The dual-process mechanism operates regardless of media literacy. Knowing that the content is propaganda does not prevent the aesthetic conditioning from running. The 145 million views came from people who knew, on some level, that they were watching Iranian state content. They watched it anyway.

Slopaganda

Philosophers Mark Alfano and Michał Klincewicz coined the term “slopaganda” in 2025 to describe AI-generated propaganda — content that is deliberately imprecise, cheaply produced, and effective not despite its roughness but partly because of it. The New Yorker’s Kyle Chayka described the Lego videos as “perhaps the world’s most potent example of slopaganda yet.”

The term captures something important about the production economics. Traditional state propaganda — the kind Russia, China, and Iran have historically deployed at scale — required significant resources: production infrastructure, distribution networks, state media amplification chains. The quality ceiling was a real constraint. Slick production looks like propaganda; rough production looks like news.

Slopaganda inverts the logic. The roughness is the point. AI-generated content doesn’t need to clear a visual quality bar to spread — it needs to clear an emotional relevance bar. The slightly-garbled text in the Epstein file label doesn’t undermine the video’s effectiveness. It may enhance it: the imperfection signals authenticity of a different kind, the kind that says “this was made by people, quickly, because they felt something.”

The production model is also genuinely cheap and fast. A team using AI generation tools can produce a three-minute animated political video in hours. They can iterate in real time — test which frame lands, which music pairing works, which narrative theme gets shared. Traditional state media cannot move at this speed. A YouTube collective can.

Viral in Enemy Territory

The paradox the Lego videos expose is not that Iranian state propaganda went viral — it’s that it went viral among American audiences.

Sharing an RT article is something American users are primed to resist. Sharing a Lego video that shows Trump bombing children is something American users on certain parts of X and TikTok do without friction, because it fits an existing emotional frame they hold about American foreign policy.

This is the exploitation of cultural fissure — not a tactic Iran invented, but one the Lego format enables at scale. The Minab school strike was real. The civilian death toll was real. The American political left had existing moral objections to the war. The Lego video didn’t create those objections. It packaged them into a shareable artifact that expressed them vividly.

The operation doesn’t need to persuade anyone. It needs to find the audience that already agrees and give that audience a form of expression that travels. The 145 million views are not evidence that 145 million people were convinced of something. They’re evidence that a sufficient number of existing skeptics found the video worth sharing — and that each share extended the network without requiring the sharer to evaluate the video as propaganda rather than as righteous content.

This is a structural advantage that conventional influence operations don’t have. Storm-1516 needs laundering infrastructure because its content has to convince people who wouldn’t otherwise agree. Slopaganda targeted at existing fractures doesn’t. The target audience’s own convictions are the distribution mechanism.

Platform Response and What It Proved

YouTube and Instagram terminated the Explosive News Team accounts in April 2026, roughly six weeks after the viral campaign began. The grounds were likely coordinated inauthentic behavior and foreign state-backed content — both standard enforcement categories.

The enforcement arrived after the content had run its course. This is not an accident of timing. It reflects a structural problem in platform content moderation: systems calibrated to detect hate speech, explicit violence, and clear policy violations are not optimized for detecting AI-generated political animation produced by a state contractor. The Lego aesthetic reduces visual similarity to known propaganda formats. The emotional framing signals satire or commentary rather than influence operation. By the time coordinated amplification patterns become statistically visible, the view count is already set.

The April 2026 termination proved the model’s timing efficiency: six weeks of operation at 145 million views before enforcement. That is a return rate any influence campaign would accept.

The Influence Tactics Breakdown

Aesthetic Frame Manipulation is the foundational technique. The Lego format encodes a specific processing instruction for viewers: engage with this in play mode. That mode deactivates the analytical faculties that would otherwise interrogate the content’s claims. The aesthetic isn’t decoration — it’s the primary operational mechanism.

Dual-Process Bypass via Sensory Stacking. The videos layer multiple conditioning signals: visual (Lego), rhythmic (trap beats), emotional (cinematic scoring), and narrative (Epstein, children’s shoes, school rubble). Each layer adds to the cumulative override of critical evaluation. By the time the political payload arrives, the viewer is neurologically primed to receive it emotionally rather than assess it logically. This is a more sophisticated version of what commercial advertisers have practiced for decades — deployed in a propaganda context at scale.

Cultural Fissure as Distribution Infrastructure. The content was targeted at existing wounds in American political discourse: the Epstein mythology, civilian casualties, imperial overreach. The virality among American audiences is not evidence of persuasion — it’s evidence of resonance with pre-existing beliefs. The distribution mechanism is the target audience’s own convictions, which removes the need for covert amplification networks. The audience does the work.

Attribution Laundering via the Grassroots Identity. “Explosive News Team” as a brand accomplishes what a GRU front accomplishes in different contexts: it places an independent-seeming face between the state sponsor and the content. The “customer” admission suggests the relationship was contractual rather than hierarchical, which adds another deniability layer. Users who shared the videos were sharing content from what appeared to be a passionate if amateur creative collective — not Iranian state media.

Slopaganda Scalability. AI generation reduces the cost of propaganda production by orders of magnitude and removes the quality ceiling as a constraint. Any state actor with access to current AI video tools can replicate this model. The barrier to entry is not production capability — it’s knowledge of the format and the target audience’s pressure points. Both are now publicly documented.

The Model Traveled

The Lego videos are not a one-off. They are a proof of concept.

When a format demonstrates a 145-million-view return with six weeks of operational life and minimal attribution cost, it becomes a template. The specific aesthetic — Lego, AI, trap beats — can be varied. What cannot be varied away is the core mechanism: playful format, emotional conditioning, cultural fissure exploitation, AI production at scale, grassroots identity cover.

Every adversarial state actor observed the Lego campaign. Every non-state influence operation observed it. Every political consulting firm operating in the slopaganda space observed it.

What Iran deployed in March 2026 against American audiences — using their own political wounds as the primary distribution mechanism — is a reproducible architecture. It doesn’t require IRGC resources. It requires AI access, audience research, and a production team willing to iterate until something spreads.

The Lego figures are incidental. The mechanism is not.

This article is part of Decipon’s Manipulation Breakdowns series, which examines specific influence operations through the Influence Tactics Protocol.

Sources: