 We're seeing more and more that disaster reliefs is becoming a big problem, global scale. More than a third of the global population has been affected by disasters. However, it is the poorest that are suffering the most. There's a growing AI for developing countries and AI for social good, so disciplined right now. What are the things that we can start using AI for? There's so many possibilities. In the cases of disaster response, it's unethical not to use robots. Unmanned aerial vehicles, unmanned ground vehicles, underwater vehicles can radically change the way we respond to disasters, the way we prevent disasters. So as part of my work with the Center for Robot Assistance Search and Rescue, we actually deploy to disasters. Most recently, the Syrian boat refugees coming into Greece. We got involved because they are actually robot lifeguards. Think of a very large life preserver, one of those throwable ones, but on a miniature jet ski robot. So now the lifeguards can zoom it out and get five to eight people can hang on and give the lifeguards time to deal with, say, the other 30 or 40 or 50 people that may be in more immediate danger. But then you discover that they don't all speak Syrian and then there's people from Afghanistan and Libya and other countries and now you're going into, we really need to start using AI to help sort out the languages. And then how do you use machine learning to start projecting? Oh, well, it's a dark night, but a full moon and the wind's down. So we should expect eight to nine boats versus two to three. That sounds like a good use of AI. I think there is a great opportunity to bring together mobile phone data that can save lives. And new ways of deploying infrastructure to be able to tackle emergency scenarios with earthquakes, hurricanes or the spread of diseases and anticipate with early cues so we can plan for it and help the people in the ground. We now have the tools to analyze the data and have much deeper understanding of what's going on. To really get much more rapidly to root causes and to give us much earlier warning on actions and remediations. Several years ago, I and a colleague got together and we looked at some opportunities for applying AI methods to social good. For years, I'd been interested in this notion that when there's a cholera epidemic in the world, you have a high rate of death that could be almost completely prevented by getting fresh water to the site in time. And so we asked the question, could we take data sets, including structure of the geography, the weather, the economy, the nature of food supplies and water supplies in a region and automatically assign a risk score for cholera and build the system that would make predictions about cholera epidemics in advance so that planners could maybe get ahead and get water to a scene, for example. And one day it lit up and said the high risk in this area of Cuba and we said, Cuba, we haven't heard of cholera in Cuba before. A few weeks later, we see a news report and we just was celebrating the fact that the system actually alerted us to this in advance and inspired us about the prospects for the future.