On February 9, Film and Television Hurricane Tim released its latest video in the early morning, focusing on the AI ​​video model Seedance 2.0 just released by ByteDance. In the video, Tim praised the generation results of the Seedance2.0 model and called it "AI that changes the video industry", including the sophistication of the video generated by the model, the movement of the camera, the continuity effect of the storyboard, and the matching of audio and video.

It was originally a routine functional experience, but due to an unexpected discovery, he bluntly said it was "terrifying".


Tim revealed in the video that when he was experiencing Seedance 2.0, he accidentally discovered that without uploading any sound files, any information or related prompts, just uploading a photo of his own face to the model, the system can automatically generate a voice that is highly similar to his own, and can even accurately match the tone characteristics.

In addition, Tim said: "The most terrifying thing is not only my voice. The photo we uploaded is only the front of the building, but the generated camera can be transferred to the other side of the building. It knows what is behind me, even if I don't tell it."

This phenomenon shocked him: It is basically certain that a large number of videos released by the film and television hurricane team in the past have been used by ByteDance for the learning and training of the Seedance 2.0 model.


He admitted frankly that he was not sure whether the relevant data authorization terms were hidden in the platform user agreement. He had never received money or been contacted for authorization. It may be legal from a legal perspective.

Tim further added that they also tested related materials from other bloggers (such as Mr. He) and found that Seedance 2.0 restored the character image to a very high degree.

He also lamented that Film and Television Hurricane was one of the first teams to upload a large number of high-definition images, sounds and other content to the cloud.

Tim said that what everyone needs to be alert to is "You can see now what will happen if all a person's data enters the AI ​​data set. It can 100% simulate any of your forms and sounds. So, can your family tell whether such content is true or false?"