Artificial intelligence may have decided you are a bad professional World War II videos or surveillance camera recordings of the Chernobyl nuclear catastrophe: Artificial intelligence (AI) is increasingly being used to create fake historical images and videos. For media researcher Roland Meyer, this trend meets a need: filling in gaps in historical events for which there are no images. ??Do you have any reporting suggestions? Send to g1 "It appears to be possible, through AI, to recombine images from the past to generate synthetic images that do not exist, but that perhaps could have existed, and thus fill in supposed or real 'gaps' in archives and historical records", he comments. "That seems like a misleading promise to me." DW's fact-checking team shows how to identify "historical" AI-created videos and how they disseminate clichés and appeal to emotions. Signs that AI is involved While it is true that much content generated by AI is easy to identify (through watermarks, notices on social media platforms or community notes on X), on the other hand, AI tools are constantly evolving. AI-generated videos thus become increasingly sophisticated, and watermarks indicating their use are often removed. Therefore, a critical analysis of images remains extremely important. In this AI-generated video, the woman's appearance is sensual and does not correspond to the typical look of women in Ancient Rome TikTok/X/YouTube via DW Some videos are obviously easy to identify, as they portray eras long before the invention of cameras. A video on TikTok, for example, that depicts Pompeii before, during and after the eruption of Vesuvius that destroyed the city is identified on YouTube as AI-generated content. When Vesuvius erupted in 79 AD, the Roman city was buried in volcanic ash and lava. The video presents typical characteristics of AI: people move like zombies, and the protagonist's appearance does not correspond to the typical look of women in Ancient Rome. Iconic moments Other videos, however, attempt to falsify moments that occurred in times when video recordings were more common. A video on X that purportedly shows surveillance camera footage from the 1986 Chernobyl nuclear catastrophe was also generated by AI. Although it does not contain a watermark, it was identified as AI-generated content in the community notes posted just below the video. This video purporting to show surveillance camera footage from Chernobyl in 1986 was also generated by AI X via DW. Another example is a YouTube video that was produced with the generative AI tool Veo. The camera zooms in on a group of soldiers holding the American flag and shows their faces as they look toward the horizon. The video is based on the iconic photograph of the raising of the US flag on the Japanese island of Iwo Jima. The photo, taken by Joe Rosenthal on February 23, 1945, became a striking symbol of the United States' military victory over Japan and one of the best-known images of World War II. The authentic photograph of Iwo Jima can be found in the United States National Archives. At the time, the scene was recorded by a cameraman, but in a static way, and not at the angle shown by the AI ??video. The photo of the US flag raising is authentic, but the video is not YouTube via DW Despite the falsification, the YouTube account that published the AI-generated video promises to feature "patriotic stories" as well as "informative and entertaining content that brings the history of the United States to life." Is AI replacing real sources? The Iwo Jima video is based on a historical event, but adds details and perspectives that have not been documented and presents all of this in a patriotic scenario. Therefore, it is useful to carefully examine the images and also who disseminates them. Where does the image come from? Who presents it? For whom? And for what purpose? Many AI-generated “historical” videos that claim to show the past fall somewhere between facts and entertainment. They can undoubtedly spark interest in historical events, but experts warn that these videos give the past a clarity and certainty that has almost never existed. "With each interpretation of a given scene, other possible interpretations and representations are eliminated, which also exist", says historian Jürgen Zimmerer. That's why it's important for viewers to recognize that they are watching a "historical" AI-generated video and view this representation with the necessary critical distance. Clichés and stereotypes Images from Ancient Rome are obviously fictional, but in the case of images from the Second World War, for example, there is a risk that a large number of AI-generated forgeries will end up replacing authentic documents. “Especially because these AI images are adapted to our expectations and our viewing habits,” observes researcher Roland Meyer. AI models use both authentic sources and created images. The datasets used for its training include historical photographs, but also digitized paintings, film scenes and video game images. "The images of the past generated by AI mainly reproduce the patterns, clichés and stereotypes by which this past is usually represented. They are, above all, images of our expectations: not a window into the past, but a mirror of our present", says Meyer. These AI creations also raise questions about liability. Unlike reenactments used in documentaries, in which directors or producers consciously decide how to represent the past, AI video footage is generated from a