I bet it doesn't take long before AI can pull data from several videos of the same event and give a nearly 3D version of it, depending on camera angles provided.
So, say that 3 or 4 people at a concert record the same song and load it to the internet. AI draws the data, and compiles and video, and audio, and through technology, render a high definition, hi fidelity file. It will be interested to see what happens.