AI-Powered Disinformation Overwhelms X Amid Iran-Conflict, Experts Warn of an End to Fact-Based Discourse

When disinformation researcher Tal Hagin asked Elon Musk’s AI chatbot Grok to fact-check a post on X claiming Iranian missiles had struck Tel Aviv, the tool’s performance was abysmal.

The video in question was originally shared to X on a Sunday by an Iranian state-run media outlet, but Grok consistently got both its filming location and date wrong. To make matters worse, the chatbot tried to defend its incorrect conclusion by pulling out an AI-generated fake image as proof.

Hagin later responded to the unhinged interaction, writing, “Now Grok is replying with AI slop of destruction. Cooked I tell you.”

This single exchange perfectly encapsulates how disconnected X has become from reality in the wake of the U.S. and Israel’s February 28 attack on Iran. As WIRED reported shortly after the strike began, the platform was quickly overwhelmed by disinformation, with accounts spreading fake and repurposed old videos to push false narratives.

As the conflict has dragged on, the flood of disinformation has only grown more severe. In recent days, the crisis has been supercharged by a wave of AI-generated images and videos, and Grok has repeatedly produced false information when asked to verify claims posted on its own platform. AI fakes exaggerating attack damage are widely shared by paid blue check-mark accounts and Iranian official accounts alike.

The widespread availability of simple, accessible AI image and video generation tools has led to far more convincing, sophisticated fake content than what was seen in past conflicts. For example, on March 2, Iranian officials and state media outlets shared AI-generated videos claiming a high-rise building in Bahrain was on fire amid the conflict. Many of these fabricated visuals look authentic enough to fool most casual viewers: one image claiming Iran shot down a U.S. B-2 bomber and detained U.S. troops garnered more than one million views before it was removed from the platform, while images purporting to show Delta Force members captured by Iranian authorities racked up more than five million views before being taken down.

Not all AI content spread on X is convincing enough to pass for real. One viral video, for instance, claimed to show Iranian forces mass-producing missiles inside an underground cave, and was obviously fabricated—yet it was still shared by dozens of accounts and amassed more than one million views.

The Iranian government is also using AI to push openly antisemitic propaganda, according to researchers from the Institute of Strategic Dialogue (ISD), who shared their exclusive analysis with WIRED. Accounts linked to a pro-regime propaganda network on X have shared AI-generated content that falsely depicts Orthodox Jews leading U.S. soldiers into war or celebrating the deaths of American service members.

Several accounts in this same network also shared a fabricated video that claimed to show a line of young girls walking past former President Donald Trump wearing only underwear. ISD reports the post gained more than 6.8 million views before it was removed, though copies continue to circulate on X via other accounts.

“What stands out about this conflict is the dramatic spike in AI-generated content I have to debunk on a daily basis,” Hagin told WIRED. “This is likely because AI is now advanced enough to fool even working journalists, and users can create this AI slop in minutes with zero consequences. The longer we go without regulations targeting AI abuse, the more harm we will see. Unless we implement policy changes now, the explosion of AI-powered fake news will push us past the point of no return for a fact-based world.”

After the wave of AI-generated fakes began overwhelming the platform last week, X announced a temporary policy: it would demonetize blue check-mark accounts that shared unlabeled AI-generated videos of the armed conflict. X did not respond to WIRED’s request for comment about how many accounts have been demonetized since the policy rolled out. Until recently, multiple Iranian official accounts paid for X’s premium subscription service, which grants accounts a blue check mark, boosted algorithmic engagement, and eligibility to earn ad revenue from posts.

Non-AI disinformation has also run rampant on X in recent days. A key false narrative focuses on the February 28 attack on a primary school in Minab, Iran, which killed more than 168 people, 110 of them children. Pro-Trump accounts have recycled footage from other parts of the current conflict to push the false claim that the Iranian government fired the missile that hit the school. A video released by an Iranian news agency Sunday, which was independently verified by The New York Times, shows a U.S. Tomahawk cruise missile striking a naval base located adjacent to the school. Despite Trump’s false claims that Iran possesses Tomahawk missiles, the U.S. is the only party to the conflict that deploys these weapons.

While the bulk of war-related disinformation, including AI-generated content, is circulating on X, Meta also faced criticism this week for its failure to handle AI fakes. On Tuesday, Meta’s independent Oversight Board released a statement criticizing the company’s approach to labeling AI-generated content, arguing that Meta’s policies “are neither robust nor comprehensive enough to handle the scale and speed of AI-generated misinformation, particularly during crises and conflicts.” Meta responded in a public statement that it welcomed the board’s findings.

“As AI-generated images and videos grow increasingly sophisticated, users rarely question visuals that are presented as ‘evidence’ to support pro-Iran narratives when they look authentic,” Isis Blachez, an analyst with media watchdog NewsGuard, told WIRED. “Even the tools we have to check a piece of content’s authenticity have major flaws. For example, AI detection tools are not consistently able to identify AI-generated content correctly.”