Robots, Pipes, Stampedes, Oh My!

Thirty two people from around the world have gathered in Bellagio to talk about the future. We are part of the Rockefeller Foundation’s “searchlight function,” an effort by the organization to work with teams around the world to uncover new and future developments that will affect the poor. In addition to about 10 sets of researchers, there are an additional 6-7 guests who have been seeded among us to add complexity and alternative perspectives to the conversation. Which is why, at breakfast yesterday, I found myself learning about the way robots can be used to find water leaks inside pipes and detect bank robberies before they happen.

Let me explain. Old leaking and cracked pipes are a big issue in many countries—developed and developing. The last month alone in Mumbai, the media has reported at least 5 major pipe bursts due to festering leaks. Smart robots can actually prevent this from happening by finding current or potential weak spots and patching them up.

As Andre-the-robot-man explained to me, the robots are programmed to sense dissonance in patterns. In the case of a pipe, they can sense when there is a break in the surface because it is different from the other parts of the pipe. The same kind of technology can also be used to identify when a bank is being robbed or when a crowd is getting out of control. The system maps the usual behavior, and when it notices a break in the pattern, it can alert the main system. Think about how useful this would be for crowd situations in India, where there are stampedes several times a year that inevitably cause unnecessary deaths of dozens, and sometimes hundreds of people. The technology developed by Andre’s company can actually detect a simmering situation about 4-6 minutes before it explodes—giving the powers that be time to react and prevent it.

At Bellagio, we are quite far from these realities--Indian stampedes and robots that find leaks--but the peacefulness here makes it surprisingly easy to focus and get on with the thinking.