You just know that video is going to be turned into training data of some kind or another. I think visual and audio versions will be around forever, and I hear that you can enter inaccurate words in the audio prompt and still get through. Might as well train it funny if you must train it at all.