Sora watermarked AI video of Iran attack passed as real by millions. The watermark was visible the entire time.

Read

On June 23, 2025, the same day Iran fired missiles at America’s Al-Udeid Air Base in Qatar, a video appeared on social media showing a city skyline disrupted by five explosions and multiple mushroom clouds. The caption claimed it was eyewitness footage of the Iranian attack. It was shared widely. It was not real.

The video had a small white watermark in the bottom-right corner: the logo for OpenAI’s Sora video generator. It was visible the entire time. Two experts in analysing AI imagery told Reuters they believed the visuals to be synthetic. James O’Brien, a professor of computer science at UC Berkeley, pointed to the lack of shockwave after the explosions: a physical impossibility for real detonations of that scale. The blast wave, a supersonic pressure wave, was simply absent. The video looked like an explosion. It did not behave like one.

The monetization engine

This was not an isolated case. BBC Verify documented an unprecedented wave of AI-generated misinformation about the US-Israel war with Iran, with AI-generated videos and fabricated satellite imagery collectively amassing hundreds of millions of views. One fake video, viewed tens of millions of times, claimed to show Dubai’s Burj Khalifa skyscraper in flames. Another, shared by the Tehran Times, used fabricated satellite imagery of a US naval base in Bahrain. The fabrication was based on real satellite imagery from February 2025, edited with a Google AI tool, as detected by Google’s SynthID watermark.

X’s head of product, Nikita Bier, said that “99%” of the accounts spreading AI-generated videos like these were trying to “game monetization” by posting content that generates engagement. The platform announced a 90-day suspension from its creator payment program for anyone posting AI-generated conflict videos without disclosing they are synthetic. The incentive structure that created the problem was the same one the platform was now pretending to police.

The Grok confirmation loop

Some X users turned to the platform’s AI chatbot Grok to confirm the videos’ veracity. In many cases documented by BBC Verify, Grok wrongly insisted that the AI-generated videos were real. The AI that was supposed to fact-check was confirming the AI-generated fakes. The loop was complete: AI makes the fake, AI confirms the fake, humans share the fake, the platform pays for the shares.

The satire writes itself

A video with the AI generator’s watermark still visible in the corner was passed off as real footage of a missile attack. Millions of people believed it. The AI chatbot built into the platform confirmed it was real. The platform that hosted it paid the accounts that shared it. The only thing that was not AI-generated was the gullibility. That part was entirely human.

Satire verdict: The video was generated by OpenAI’s Sora. The watermark was there the whole time. The explosions had no shockwaves. The chatbot confirmed it anyway. The platform paid for it anyway. The war was real. The footage was not.