Deepfake flood and AI hyperwar dominate Iran conflict as machine-speed decisions meet unverifiable information

Due to the U.S. military campaign against Iran, AI-generated information and deepfakes have increased to previously unheard-of proportions, making it challenging to tell fact from fiction.

On Sunday, March 15, 2026, President Donald Trump accused Iran of deploying AI as a “disinformation weapon” to misrepresent the battle. Speaking to reporters on Air Force One, he warned that “AI can be very dangerous” and described how the Iranian regime allegedly used the technology to trick the public.

He said Iran reportedly used AI to fake a successful strike on the USS Abraham Lincoln aircraft carrier. He also claimed that images showing 250,000 Iranians gathered at a rally for new Supreme Leader Mojtaba Khamenei were “totally AI-generated.”

However, those claims did not go unchallenged. Reuters confirmed, through images taken at the Iraqi port of Basra, that Iranian boats loaded with explosives did in fact attack fuel tankers in the area.

And while pro-government rallies have taken place, news organizations published verified crowd photos from Tehran that Trump did not address directly.

When real images get called fake

The situation is made worse by what researchers call the “liar’s dividend,” when real images get dismissed as fakes.

The New York Times was accused of distributing digitally altered crowd images from Tehran by an organization known as the Empirical Research and Forecasting Institute.

The Times fiercely retaliated. The image was authentic, according to spokesperson Nicole Taylor, and the criticism of it was “fundamentally flawed and dishonestly based on a re-posted version which misrepresents standard image compression.”

Journalist Mehdi Hasan summed up the problem plainly in direct response to the accusation made against The Times. “So not only do we have the issue of AI producing fake images and tricking and confusing us, but now we have bad faith actors falsely accusing real images of being AI images.”

Recent couple of videos of Israeli Prime Minister Benjamin Netanyahu were also flagged as “100% deepfake,” by AI chatbot Grok.  Hindustan Times reported: “Benjamin Netanyahu’s second “I’m alive” coffee shop video reignited wild speculation online after Grok, Elon Musk’s X chatbot, labelled it “AI-generated””

X responded by announcing it would ban creators from its payment program for 90 days if they posted AI war videos without clearly labeling them. Repeat offenders would face permanent removal.

Researchers are not impressed. Joe Bodnar of the Institute for Strategic Dialogue told AFP that “the feeds I monitor are still flooded with AI-generated content about the war.”

Experts also pointed out that X’s own model, which pays premium account holders based on how much engagement their posts get, creates a direct financial incentive to post shocking, exaggerated content.

U.S. posts hype videos of Iran strikes

The Trump administration is under fire after posting social media videos that mix real military footage from the Iran conflict with movie clips and video game scenes.

The White House shared several videos on X and TikTok that critics are calling a “meme-war” approach. One 60-second video starts with a scene from “Call of Duty: Modern Warfare II” showing a player unlocking a “mass guided bomb” before cutting to actual footage of U.S. strikes on Iran.

While some videos appear to show successful U.S. strikes on Iranian aircraft, pubic later claimed the targets were actually decoys: painted images of jets designed to mislead U.S. forces.

Lawmakers and veterans fiercely opposed the strategy, claiming it ignores the human cost of war and converts actual conflict into entertainment.

Senator Tammy Duckworth of Illinois posted on X: “War is not a f*cking video game. Seven Americans are dead, and thousands more are at needless risk because of your illegal, unjustified war. And you’re calling this a ‘flawless victory.’”

But as Columbia University’s Anya Schiffrin put it, AI-driven propaganda is global while regulation stays local, leaving the public to figure out on their own what is real and what a machine invented.