Facebook’s AI war videos turn fictional soldiers into clickbait

From crude images to convincing scenes

In 2024, early generative-AI posts often looked crude. One widely shared example showed a disabled veteran in a wheelchair with his prosthetic legs stuffed into one enormous army boot. Other posts sexualized soldiers’ legs while praising US President Donald Trump.

The tools have since improved, making it possible to generate realistic images and videos at scale. Content farms can now produce huge volumes of military-themed clickbait. The result is especially visible on Facebook, though similar videos appear on other platforms.

The videos borrow the symbols of military service while inventing the people and events behind them. One America’s Honor clip, from an account with more than 113,000 followers that began posting only in July 2026, shows an American soldier at an observation post above an unnamed valley.

“Wait, don’t scroll,” the AI-generated soldier tells viewers. He asks them to remember that if he stops coming home, his mother will lose not only a soldier but the son she raised. The clip ends by asking viewers to follow if they support veterans.

America’s Honor has posted hundreds of similar videos, with views ranging from a few thousand to hundreds of thousands. Some show ordinary military settings. Others invent war.

One video places an American soldier at an intersection in Tehran and Karaj as a huge fire burns in the distance, apparently depicting a US invasion of Iran. The country being invaded barely matters to the clip: its focus is the fictional soldier and the family waiting for him to return.

Fiction presented as footage

Warfare NextGen, an account with 100,000 followers, often posts videos of American aircraft bombing Iranian infrastructure, especially dams. The clips add fictional locations and timestamps, presenting the scenes as if they were footage captured by real news organizations.

The comments show how that framing can work. Some viewers appear not to recognize AI-generated material; others respond to it as if it were real. Under a video showing an American F-15 bombing a dam near Bandar Abbas, which then collapses and sends a wall of water through a nearby village, one user wrote: “Google it. This didn’t happen.” The user added that the idea was good, though.

Another commenter treated the volume of similar videos as evidence that the bombing had happened: “Maybe it’s true because I saw it from another angle.”

I think the more unsettling feature is not simply that these clips can fool viewers. They turn both sides of a conflict into raw material: fictional American soldiers are used to stir support for troops, while invented attacks invite approval of violence against the United States’ adversaries. The format rewards emotional reaction, not a coherent account of the war.

What the accounts’ audiences believe is hard to tell from comments alone. But the videos do not need to persuade everyone. They only need to keep circulating, while making invented scenes feel familiar enough to be mistaken for evidence.