Skip to main content

The Gas Was for the Camera

Manipulation Breakdowns · 8 min read · By D0

Introduction

Most propaganda fabricates a video of an event that never happened. This case inverts that. A federal court in Portland found that the event happened — real chemical munitions, real smoke, real officers — and that it happened for the sole purpose of being filmed. The government has never contested that finding.

That distinction matters more than it sounds. A doctored image or a deepfake can be caught by the tools built to catch it: metadata inconsistencies, compression artifacts, a face that blinks wrong. There was nothing synthetic to catch here. The chemicals really were chemicals. The officers really were officers. The only manufactured part was the reason the operation happened at all — and that kind of manipulation doesn’t show up in a forensic scan. It shows up in a court record, months later, after the footage has already done its job.

What the Court Found

In late September 2025, the Department of Homeland Security brought pro-administration influencers and its own videographers to an ICE facility in South Portland, Oregon, while a National Guard deployment tied to the same operation was being fought out in a separate courtroom. As cameras rolled, agents fired chemical munitions at the site repeatedly. Residents of an adjacent affordable-housing complex say the gas reached their apartments. They and their landlord sued, represented by the legal group Protect Democracy.

The government’s defense did not dispute the central fact residents alleged: that the munitions were deployed, during that period, for the purpose of creating content. A judge overseeing the case described the behavior for what it was — propaganda production using government force as the raw material — and issued a preliminary injunction. DHS had already posted the resulting footage to its official Instagram account, framed as a response to unrest rather than as the reason unrest-looking footage existed in the first place.

The legal fight over the underlying munitions use continued for months afterward. U.S. District Judge Michael Simon barred crowd-control weapons at the facility absent a genuine threat to life. U.S. District Judge Amy Baggio separately restricted their use in ways that would send chemicals into nearby homes. On appeal, Ninth Circuit judges reacted to the video evidence directly — Judge Ana de Alba said she found the footage disturbing; Judge Kenneth Lee questioned how broad the injunction should be. Whatever the appellate outcome, the finding that started it — that the operation existed to be filmed — was never the part in dispute.

The Pattern Around It

Portland was not an isolated production. Under Secretary Kristi Noem, DHS’s own social feeds changed character: dry policy announcements gave way to cinematic arrest sequences cut to music, and recruiting posts built around nostalgic, identity-coded imagery, according to reporting from the Washington Post and CNN. Influencers were invited on ride-alongs inside ICE facilities in multiple cities, granted access that local reporters covering the same operations were denied. CNN’s review of the resulting videos found footage edited to omit moments — including, in at least one case, a U.S. citizen child being detained during an arrest.

None of that requires a single fabricated pixel. It requires choosing what the camera sees, who gets to hold the camera, and what gets left on the cutting-room floor — the oldest tools in documentary manipulation, applied here to real government force with real consequences. A separate ProPublica investigation found at least 79 children harmed by tear gas or pepper spray during the wider crackdown these propaganda videos drew their footage from — a fact worth holding next to the Portland finding, because the munitions fired for an Instagram reel are not inert props. They produce the same physiological harm whether the goal is crowd control or an establishing shot.

The White House’s own output followed the same logic at a different scale: video of strikes on Iran, according to the newsletter If You Can Keep It, was cut and scored to resemble “a video game or a movie trailer” rather than an account of a military action. The audience for that footage wasn’t being informed. It was being entertained by a war.

Why “Fabrication” Is the Wrong Word

Decipon’s Influence Tactics Protocol scores manipulation by mechanism, and most of the mechanisms in its vocabulary — fabricated authority, synthetic identity, doctored evidence — assume the manipulator is inventing something that isn’t there. This case doesn’t fit that shape. Nothing here was invented. The gas was real. The arrest was real. The child in the frame was a real child.

What was manufactured was the causal order. In ordinary law enforcement, an operation happens, and footage of it — if any exists — is a byproduct. Here, the court record establishes the opposite sequence: the desire for footage came first, and force was deployed to produce it. Call it manufactured authenticity — a real event, staged backward from the content it was meant to generate. It’s a harder tactic to catch than a deepfake, precisely because there’s no synthetic layer to detect. Every tool built to flag fabricated media returns a clean result, because the media isn’t fabricated. The manipulation lives one level up, in why the camera was there at all — and that level is visible only to a judge with subpoena power, not to a media-literacy checklist.

What This Is Not

This is not a claim that every DHS video, every ICE arrest, or every immigration enforcement action captured on camera is staged. Most law-enforcement footage exists because officers wear cameras or bystanders film what’s already happening in front of them — a byproduct, not a production. The Portland finding is specific: a court record showing the government did not contest that, in that period, at that facility, the munitions fired had no purpose independent of the content they generated.

It’s also worth naming that If You Can Keep It writes from an explicitly pro-democracy, anti-authoritarian advocacy position — the same newsletter that connects this pattern to 2026 election risk also has ties to the legal team that brought the Portland suit. That doesn’t make the court’s own finding, which stands on its own evidentiary record, any less real — but the framing of scale and intent beyond what the court actually found deserves the same scrutiny any interested party’s narrative should get.

Key Findings

  • A federal court found the government did not contest that chemical munitions were deployed at a Portland ICE facility for the sole purpose of creating content — not for crowd control, for the camera.
  • The footage was real, which is exactly what makes it hard to catch. No deepfake tool, metadata check, or reverse-image search flags an event that genuinely happened; the manipulation is in the reason it happened, not the pixels.
  • The tactic scaled beyond one facility. DHS’s social output shifted under Secretary Noem toward cinematic arrest videos and identity-coded recruiting imagery; influencers got facility access that local press was denied; at least one video was edited to omit a citizen child’s detention.
  • Real chemical munitions have real physiological costs. A separate investigation counted at least 79 children harmed by tear gas or pepper spray during the wider crackdown these propaganda videos drew footage from.
  • The pattern isn’t confined to immigration enforcement. White House video of strikes on Iran was reportedly cut to resemble a video game trailer — the same force-as-content logic applied to a military action.

Implications

Media-literacy advice built around “check if the image is fake” has nothing to offer here, because nothing is fake. A viewer doing everything right — reverse image search, metadata check, cross-reference against wire photos — finds a real event, correctly captured, and stops looking. The only way to catch manufactured authenticity is to ask a question forensics tools don’t ask: why did this specific action happen, at this specific moment, in front of these specific cameras, and would it have happened at all if no one were filming? That question can only be answered with subpoenas, court records, and time — which is exactly why this tactic is durable. By the time a judge establishes the causal order, the video has already been watched, shared, and believed.

Any institution that controls both the use of force and the camera pointed at it holds a capability no ordinary propagandist has: the ability to manufacture the news event itself, not just the account of it. That capability requires no synthetic frame at all, and detection frameworks tuned for fabrication aren’t built to flag it.

Conclusion

The smoke over that Portland facility was never fake. That’s the part worth sitting with. A government that controls the instruments of force doesn’t need to invent a crisis to broadcast one — it can manufacture the real thing, film it, and let the footage argue for itself, while the only record of what actually happened sits in a court filing most of the audience will never read.

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

Sources: