Iranian security chief Ali Larijani was killed in an Israeli attack on March 17, but a video purporting to show the attack has been circulating online that was actually generated by AI. The clip was posted by an account that primarily posts synthetically created visuals of the joint US-Israel war against Iran that began on February 28th.
A caption in Hindi on a Facebook video shared on March 31, 2026, reads: “Iran is on its last stand, as Israel is selectively killing. This attack will surely scare you! Iran’s head of national security, Ali Larijani, was killed in this attack. The intensity of the attack shook the surrounding Iranian soil.”
The video appears to show a series of airstrikes hitting the building.
Screenshot of a false post taken on March 6, 2026, with the AI symbol and red X added by AFP
The video was also circulated in similar Facebook and X posts (archive link) as countries tried to end more than five weeks of fighting sparked by U.S. and Israeli attacks on Iran, effectively closing the strategic Strait of Hormuz and responding by attacking targets across the Gulf. On April 7, Iran rejected a proposed ceasefire with the United States and Israel.
Since the start of the war, US and Israeli attacks on Iran have killed Supreme Leader Ayatollah Khamenei and the Islamic Republic’s entire political and military elite (archive link).
Iran’s Supreme Council for National Security confirmed his death on March 17, after Israel announced that it had killed him in an airstrike (archive link).
However, the circulating clip does not depict the airstrike.
A reverse image search on Google using keyframes from the accidentally shared video turned up the same clip posted on Instagram a few weeks ago (archive link).
The March 10 video was also shared with a disclaimer that it was AI-generated.
Comparison of the accidentally shared video (L) and a screenshot of the Instagram video from March 10th. AI symbol and AI disclaimer added by AFP are highlighted
The video was shared by an account that regularly posts AI-generated visuals of Middle East wars.
The user of the account, who did not want to be named, told AFP on April 3 that the long caption on the post describing the scene was an AI prompt used to generate the video with Open AI’s now-closed Sora 2 video generation tool (archived link).
The user also said their purpose was “to share for entertainment, not to spread fake news.”
Upon further analysis of the video, we found that it contained visual inconsistencies, a tell-tale sign of synthetic content.
Despite being hit multiple times by missiles, the surrounding buildings in the video are not shaken or damaged. Despite the explosion, there were no signs of panic and people could be seen walking calmly down the street.
Screenshot of a video that was accidentally shared. AFP highlights an undamaged building and a person walking calmly during the explosion.
Additionally, analysis by the Hive Moderation AI detection tool found that the video “may contain AI-generated content or deepfake content” (archive link).
Screenshot of video analysis from Hive Moderation AI detection tool
AFP here debunks other misleading claims about the Middle East wars.
