You first need a target video to use as the basis of the deepfake and then a collection of video clips of the person you want to insert in the target. There are several methods for creating deepfakes, but the most common relies on the use of deep neural networks that employ a face-swapping technique. "You have a lot of AI assistance with CGI, but at the end of the day there is a human with a human viewpoint controlling what the output is going to be," López said. Note: Computer-assisted technologies like Photoshop and CGI are commonly used to create media, but the difference is that humans are involved in every step of the process, barring recent development's like Adobe's generative Firefly tool. When it comes to deepfakes, the user only gets to decide at the very end of the generation process if what was created is what they want or not outside of tailoring training data and saying "yes" or "no" to what the computer generates after the fact, they don't have any say in how the computer chooses to make it. (Images like these, when combined with misleading information, are commonly referred to as "shallowfakes.") What separates a deepfake is the element of human input. The AI-generated pope in a puffer jacket, or the fake scenes of Donald Trump being arrested that circulated shortly before his indictment, are AI-generated, but they're not deepfakes.
How are deepfakes different from other kinds of manipulated media?ĭeepfakes aren't just any fake or misleading images.
0 kommentar(er)
