I’m working on a movie where I want full control over the movement. I’m using

Blender and CozyClay for the animation pipeline. But I’ve been noticing that many AI video models only need an image reference to create convincing motion. That raises a question for me: are we moving toward a workflow where directing movement is less about animating every detail and more about choosing the right reference, intention and edit? I’m interested in keeping the human hand in the loop, especially when movement carries emotion or story. For those working with AI video: how much control do you want over movement? Do you prefer directing it frame by frame, or are you okay with letting the model interpret the motion?