Elon Musk’s X Finally Tries to Stop the Epidemic of AI-Generated War Footage

The social media platform X has been flooded with fake photos and videos ever since President Donald Trump launched a new war on Iran last week. But X’s head of product Nikita Bier announced a new policy Tuesday that he hopes will disincentivize accounts from sharing AI-generated fakes. At least when the motivation for sharing is purely financial.

“Starting now, users who post AI-generated videos of an armed conflict—without adding a disclosure that it was made with AI—will be suspended from Creator Revenue Sharing for 90 days,” Bier wrote Tuesday in a post on X.

“Subsequent violations will result in a permanent suspension from the program. This will be flagged to us by any post with a Community Note or if the content contains meta data (or other signals) from generative AI tools,” Bier continued.

It’s not immediately clear whether there will be requirements for how large a disclosure may need to appear and whether it needs to be embedded into the video or can be merely included in the text of a tweet. There are plenty of loopholes that X accounts use for impersonation, like making a username so long that the word “parody” only appears if you click through to view a given account’s profile. The potential loopholes here also seem endless.

U.S.-Iran War fakery

Fake photos and videos have gotten millions of views in recent days, ever since the U.S. and Israel launched a war in Iran that has killed Supreme Leader Ayatollah Ali Khamenei and a large number of other Iranian officials. And until Tuesday it showed no signs of slowing down.

One of the fake images included a U.S. fighter pilot who was shot down and supposedly mistaken for an Iranian by a Kuwaiti man with a pipe. The image includes the SynthID watermark from Google, meaning it was created using one of Google’s generative AI products.

If you look closely, the people in the image are also missing some fingers, a classic sign of fakery by AI.

Several different accounts on X have also shared footage that purports to show Tel Aviv, Israel, getting bombarded with rockets.

But the video has several big red flags that indicate it’s been generated with AI, according to BBC disinformation tracker Shayan Sardarizadeh. The most glaring might be the cars on the street, which are in bizarre shapes and don’t look like real cars.

But there’s also the audio, which includes someone off-camera saying “Tel Aviv, I can’t believe this,” in an unnatural way that’s just a little too perfect if you’re trying to spread fake information about a specific location.

Many X users have asked xAI’s Grok whether the video is real and it seems to be consistently responding that it is. One user who shared the video even insisted that it must be real because Grok said so.

#ULTIMAHORA
🚨 CASTIGA IRÁN CON FURIA A TEL AVIV

Y si, si es real, si gustan chequen con @grok… pic.twitter.com/oypZhe0OrE

— La Catrina Norteña (@catrina_nortena) March 3, 2026

But Grok is an awful fact-checker and can’t be relied upon to tell you whether a video is real, just as it shouldn’t be used for anything involving World War II history. This is MechaHitler himself, after all.

Mislabeled videos

One question that hasn’t been answered by X is whether misleading images and videos that aren’t necessarily created with AI will be demonetized. Because there are plenty of other ways to mislead people on social media in a time of war.

A popular fake video that’s gone viral also purported to show the U.S. embassy in Saudi Arabia going up in flames. The embassy was indeed hit by two Iranian drones on Monday, according to the New York Times, but that’s not what’s depicted in the video.

BREAKING 🚨

EXPLOSIONS REPORTED AT THE US EMBASSY IN RIYADH, SAUDI ARABIA pic.twitter.com/eMQVdDICol

— Iran Times (@IranTimes9) March 3, 2026

 

In reality, the video above is about a month old at minimum, having been posted to YouTube on Feb. 6. It has nothing to do with the current war. While the video appears to be real, it’s being misrepresented as something that happened recently.

Another video was captioned “An Iranian plane VS a US ship. I can watch this all day,” racking up over seven million views. It’s actually footage from a video game. Is creating a video clip of game footage and presenting it as current events going to qualify X users for demonetization? There’s no sign that it was created using AI, which is the only thing Bier mentioned in his tweet Tuesday.

An Iranian plane VS a US ship.

I can watch this all day 😂 pic.twitter.com/qjLH2rA8V1

— Joel Fischer 🇺🇸 (@realJoelFischer) March 1, 2026

Still another video that gained significant attention supposedly showed the “CIA headquarters” in Dubai with smoke billowing out of it after being hit by Iran. An account that frequently posts disinformation falsely claimed that authorities in the UAE were arresting anyone who shared the footage. But the footage is actually from 2015.

Footage of CIA headquarters in Dubai targeted this morning by Iran, have emerged.

UAE is arresting those releasing the footage. But they can’t hide this. pic.twitter.com/K3iPuvH4WU

— Syrian Girl (@Partisangirl) March 1, 2026

Supposedly respectable people can get caught up in sharing these fake images and videos, as we’ve seen repeatedly in just the past few days. Fox News host Bret Baier shared the fake embassy video and Texas Gov. Greg Abbott shared the video game footage, though both deleted their tweets.

Hotbed of disinformation

X is ground zero for fake photos and videos whenever news breaks, whether about war or any other topic. Elon Musk helped make the disinformation problem worse after he bought Twitter in late 2022, inviting back conspiracy theorists who had previously been banned on the site and stripping so-called legacy verification checkmarks that helped give users a sense of who they could trust. And then generative AI technology made the problem even worse.

Musk also introduced the creator revenue sharing program that created incentives for users to get the most attention, whether something was true or not. And Musk himself often shares things that aren’t true or are AI, like video of Sydney Sweeney to promote Grok.

Musk hasn’t weighed in personally on the new policy to demonetize accounts that share AI content without disclosure.

Musk’s own incentives

Far-right commentator Matt Walsh thought the new X policy didn’t go far enough. “Why not suspend anyone who shares any AI content without disclosing that it’s AI?” Walsh asked in a tweet Tuesday.

The answer is likely because Musk envisions a world where everything consumers watch on X is AI-generated. The billionaire has said as much in several discussions, including with podcaster Joe Rogan late last year.

“Most of what people consume in five or six years, maybe sooner than that, will be just AI-generated content. So music, videos…” Musk said while trailing off.

What about the political actors sharing fakes?

It seems like a good step to demonetize accounts that are sharing AI footage on X that’s not appropriately labeled, even if it’s a very modest move. But that will only disincentivize users who are sharing because they’re trying to rack up views for financial gain.

What about accounts that are sharing for different reasons, like trying to influence public opinion or betting markets outside of X’s control? Or what if they’re sharing just to stir shit? Because in the age of AI, there’s basically no hurdle for creating an endless supply of fake content, as Bier seemed to acknowledge on Tuesday.

“During times of war, it is critical that people have access to authentic information on the ground. With today’s AI technologies, it is trivial to create content that can mislead people,” Bier tweeted.

X didn’t respond to questions emailed Tuesday, which isn’t a surprise. The company has a history of being hostile to journalists since Musk purchased Twitter. Gizmodo will update this article if we hear back.