I recently saw this story about Adobe creating a new product using AI to smooth out dance movement in videos.
I can definitely see the value in something like this in the Covid era. If you have ever tried to synch up videos of people singing the same song in different rooms recorded on devices of varying quality, you know what a challenge that can be. There can be a similar benefit for dance groups that have their members recording videos in disparate locations.
But the same technology can be used to make people look like better dancers than they actually are. The Adobe researcher in the video accompany the article, I assume he is Jimei Yang, says he started working on the AI because his daughter felt his dancing wasn’t up to the standard of the guy in another video she was watching. So there is no pretense about the technology being helpful in stitching videos together or being used to analyze your movements so you can improve your skill as a dancer. It is all about making you look like a better dancer than you are.
If you do any reading about the controversies over using auto-tune to make people sound pitch perfect, you’ll find that some feel tools like these diminish the value of hard work to cultivate your skill. Others will say that it provides new options for creativity that didn’t exist before. Then there are others that won’t say anything because they depend on sounding pitch perfect for their livelihood.
One thing that will likely keep tools like the one Adobe is developing from being used as widely as auto tune as a substitute for skill is live performance. When someone is presenting a live concert, it is easy enough to lip synch to a recorded track or have vocals processed before being transmitted without being detected. Short of implants that allow an AI to control your movement, it is tough to enhance dance skills beyond your actual ability during a live performance.