What if a human combines AI-generated video and audio before uploading? Is that too much human editing?

@JonasVollmer could you go into more detail about what you mean?

@RemNi If I make a silent video with 100% of frames AI-generated, and I add a soundtrack that's 100% AI-generated, combining those into a single video file before uploading technically is an 'editing process involving a human', which is prohibited according to your resolution criteria. But I think you probably want to allow that.

@JonasVollmer If a language model is used to generate a video description and an audio description from some input prompt, and then text-to-video and text-to-audio models are used to generate the corresponding video and audio, the resulting video+audio is acceptable since the "AI" is the whole system in this case.

If a human separately generates multiple video and audio samples, and then selects which ones to combine together, then this is not acceptable, because they are adding human authorship to the process.

So if the editing process is done my an AI agent, it still counts?

Is the only human input allowed prompting, and pressing the upload button?

@soweliSon yes that would still count. Yes, the human is only allowed to do text prompting and pressing the upload button

@RemNi Does the whole process have to be directed by a single prompt, or can humans come up with multiple prompts for the various sub-tasks?

@soweliSon The whole process must be directed by a single prompt. However, it's permissible to use a complex text prompt that contains a description of what to include in each scene of the video. Text prompts that are simply a list of instructions on how to draw each frame are prohibited, e.g. :

"On frame 232, draw a line from (304, 266) to (823, 480)..."

I'll update the description, thanks for the input!