Fabricated videos purporting to show the recent high-profile heist at France's Louvre Museum are circulating widely on Hong Kong social media platforms, despite being confirmed as AI-generated content. The fake footage, which claims to document the October 19 jewelry theft from the museum's Apollo Gallery, has been shared by accounts with hundreds of thousands of followers.
The actual heist made international headlines when thieves conducted a daring daytime robbery, parking a truck with an extendable ladder beneath the museum's Apollo Gallery. The criminals used cutting equipment to break through a window and display cases to steal valuable jewelry from the world-famous Paris museum.
A Facebook reel featuring what appeared to be security footage of the theft was posted by a Hong Kong-based user with more than 240,000 followers. The 41-second video includes Mandarin-language narration stating, "At around 9:30 a.m., four masked men broke into the building from the balcony and took away the jewelry on display," while showing two separate clips of masked individuals breaking glass display cases.
However, digital forensics experts and independent verification have conclusively determined that these videos are not genuine and were likely created using artificial intelligence technology. The fabricated footage has spread beyond Facebook, appearing on other major Chinese social media platforms including Douyin and RedNote, where multiple versions continue to circulate.
AI Forensics, a European non-profit organization specializing in digital manipulation detection, conducted a detailed analysis of the suspicious videos. Researcher Natalia Stanusch explained that the footage contained multiple telltale signs of AI generation, including objects mysteriously appearing and disappearing in the hands of the supposed thieves, morphing hand shapes, and fingers that vanish mid-action.
Perhaps most damning, both videos contained partially obscured watermarks from OpenAI's Sora, an advanced AI video-generation tool. These watermarks, though attempts were made to hide them, provided clear evidence that the content was artificially created rather than captured by actual security cameras.
Independent verification by comparing the fake footage to official images of the Louvre's Apollo Gallery revealed significant discrepancies. In the first fabricated clip, the appearance of the gallery entrance differs notably from the authentic layout shown on the Louvre's official website. The second clip shows roof structures and wall artwork that do not match the actual gallery's distinctive architectural features and decorative elements.
This incident highlights the growing challenge of AI-generated misinformation spreading across social media platforms, particularly as generative AI technology becomes more sophisticated and accessible. The case demonstrates how quickly false content can gain traction online, especially when tied to legitimate news events that capture public attention.
The spread of these fabricated videos underscores the importance of media literacy and fact-checking in the digital age, as users increasingly encounter content that may appear authentic but has been artificially created. Social media platforms and users alike face mounting pressure to develop better methods for identifying and preventing the circulation of AI-generated misinformation.
