Dancing Into Different Worlds

The storm of change will mean that the nerds who used to build the computer software of old, by typing thousands of lines of code into a black box on a flat screen, might need to learn to dance.

Yeah, if you watch someone play in Virtual Reality you might think they are dancing, but that isn't what we are thinking about. There's a new type of computer science underway: one that uses choreography to train the AIs that control autonomous cars, robots, virtual beings, and even present humans in Augmented and Virtual Reality with interfaces that better serve us.

Machines can be cold. In the worst of cases, they can crush us on factory floors, or kill us in the streets, like a computer-controlled car developed by Uber did one unfortunate night in Arizona. In the best of cases, they can give us superpowers and make our lives easier, but even then they often can be made to better serve humans in a dance, of kinds.

Did you know there is an entire conference for research on these new ways of training machines called "Choreographic Interfaces?" We didn't either, until we were introduced to Catie Cuan, who is currently studying for a PhD at Stanford University on the topic. As she showed us around Stanford University's AI and robotics labs, where she is working on her studies, we discussed how innovation teams will soon be changed by her, and other people who are using techniques that humans use in dance choreography to train computers.

Photo credit: Robert Scoble. Catie Cuan, Stanford AI student, tells us about using dance choreography to train robots in the lobby of the Bill Gates building at Stanford University, which is the building where Google started.

What makes her unique among most of the computer scientists we've met? She's a former dancer. "Dance and choreography is all about moving with intention," she says, as pointing out that she spent more than two decades dancing and that she uses the knowledge she gained by studying how human bodies could be made to move in different ways, and then translates that training and knowledge to teaching robots and virtual beings to move.

Cuan is working for automobile companies who are developing autonomous cars, among others, as part of her studies (she did some of her schooling at the University of Illinois, which is where Tesla and PayPal started, among others). She told us that humans go through an intricate dance of sorts as they wave each other through intersections, for instance, and cars without human drivers can't communicate with humans on the street or in other cars that way, so they need to be trained both how to recognize human gestures, say, like a police officer standing in the middle of an intersection directing traffic, and communicate with humans their intention to move or stop, but do so in a pleasant, human way.

Now, as computing will be everywhere, we need new interfaces. Ones that respect us, understand us, warn us properly, all in a very human way. Building these new interfaces requires new kinds of innovation teams with, even, a dancer or two to help out. Why? Well, she says that she sees an optimistic future for human/machine interactions, and what better way to make the machines more fun and engaging to be around? As this storm of change comes, though, it may frustrate many people who want things to stay the way they are, or who can't wrap their heads around the fact that they need to work with someone doing choreographic work. Many pioneers face similar resistance in their careers. We suggest that if you are feeling yourself resisting this change, you might need to be the one to change lest you, or your company, be left behind.