AI-Generated Fake Videos on Iran War Go Viral with Millions of Views

A flood of fake videos and images created by artificial intelligence has overtaken social media in the early weeks of the Iran war, further complicating the information landscape surrounding the conflict. The New York Times identified over 110 unique AI-generated pieces of content in the past 14 days, including scenes of devastating explosions in Tel Aviv that never happened and troops protesting on nonexistent streets. These fakes, rapidly spreading across platforms like X, TikTok, and Facebook, have already amassed tens of millions of views, adding a layer of confusion to the online debate.

The issue isn't new but has reached unprecedented scale in this war, the first of significant magnitude in the era of publicly accessible AI tools. For instance, a fake video of Iranian missiles hitting Tel Aviv circulated in over 300 posts, while another showing the Burj Khalifa skyscraper in flames in Dubai garnered millions of views before being taken down. It goes further: images of American Delta Force special forces being captured by Iranians reached 5 million views, according to analysis by WIRED. But what motivates this production? Part of it stems from Iranian state propaganda, which uses AI to project fictitious victories and distort real losses, as highlighted by the Institute for Strategic Dialogue in a recent report.

Meanwhile, independent creators profit from engagement, monetizing views and ads on sensationalist content. Tools like Sora and Runway facilitate the rapid creation of these materials, and even xAI's Grok mistakenly validated some as real in automated responses. In practice, this erodes trust in online videos, especially in a conflict where misinformation can influence public opinions and political decisions.

How to Identify and Combat Fakes

• Look for watermarks or visual errors, such as extra fingers on hands or inconsistent shadows.
• Verify sources: use sites like BBC Verify or CNN Fact Check to check origins.
• Check timestamps: many fakes reuse old clips from other events.

Platforms like X have already deleted thousands of posts, but the volume continues to grow. American and Israeli authorities warn of the risk of escalation through false narratives, and Europol predicts that by the end of 2026, 90% of web content could be AI-generated. Looking ahead, tech companies like Meta and Google promise better detection tools, with updates expected in April that use AI for real-time deepfake flagging.