Sounds like maybe an application for geometric Ai, which operates on 3D objects, rather than on 2D pictures. I once saw a video of dolphins interacting with “Bubble rings”, where an acoustic machine generating these, a bit like underwater smoke rings, this machine blowing them out periodically was placed on the sea bed underneath some wild dolphins. At first it provided much amusement with the dolphins shooting through them. Then, astonishment, within a few minutes, the dolphins started mimicing the machine, producing the rings themselves, emitting them from their breathing glands.