 2.5 exabytes of data are generated every day. This is equivalent to 90 years of high definition video. Humans alone cannot hope to process the volumes of data necessary to make informed decisions. We need assistance. We need artificial intelligence. The relationship and partnering between human and technology is increasingly important. Until now, humans have been the masters and technology the slaves. This needs to change. Today's AI systems are capable of achieving complex goals in an agile and flexible ways. However, while these systems are good at solving narrowly defined tasks, they don't know how to collaborate with a human or how to act as part of a problem-solving team. Increased processing power provides the opportunities to rectify this and has led to new modes of human-computer collaboration. My big idea is for humans and AI systems that I call software agents to work together. They have complementary expertise that can be harnessed to tackle difficult problems. I call such systems human agent collectives and my research at Imperial College London explores the science and the engineering of such systems. Our agents use Bayesian mathematics and machine learning to determine when to ask a human or another AI system for assistance and how to respond when their plans are modified or overridden. Our agents use distributed optimization technologies to determine who in the team is best placed to solve which task. Human agent collectives can be used for a variety of applications. Today, I'm going to focus on disaster response, where our technologies have been used in Haiti and Nepal. In particular, we've worked closely with Rescue Global, an international crisis response charity. Our agents summarize huge volumes of information in order to identify potential casualties. This information is a mix of text and photos and sensor readings. It's constantly changing and often incomplete. Human first responders on the ground can add trusted observations and the agents will rework their assessments to be consistent. First responders are adept at analyzing emerging views of a situation to identify areas of high uncertainty. They can then task the agents to manage the humans who have volunteered to help from afar. The agents can learn which humans are good at this task and place greater reliance on them. As the situation becomes clearer, the first responders act. Using their experience and intuition, they determine which response teams are best suited to which areas. The agents can then construct a schedule that minimizes completion time. The first responders can then either accept or tweak the plan as necessary. As information continues to come in, they jointly monitor the ongoing operation. The agents monitor incoming information streams to determine things that might change priorities. The humans monitor the response teams for unexpected difficulties or for delays. During the aftermath of the Nepal earthquake, Rescue Global used human agent collective technologies to determine the placement of water filters around Kathmandu. Using social media, satellite imagery and our smart agents, they identified 12 areas of need, four of which were previously unknown. Disaster response is but one application of these technologies and many of the partnerships in these systems will be of the type I've outlined. Others, however, will differ. In some cases, the AI will direct the humans, as happens with Uber drivers today. In other cases, the AI will replace the humans, as has happened with financial traders. This will affect highly skilled, highly qualified workers. However, I believe that there will always be tasks that humans are best at, from social work to science. These involve intuition, experience and dexterity. Human and computer partnerships of the type I've outlined complement each other. They complement each other's strengths and weaknesses. This leads to a rise in the humans as well as in the robots. However, for this to happen, the machines need to move from being frustrating tools to active partners that can help us. AI is here today and it will become ever more important as we generate ever more information about everything. AI can enhance our lives and working collaboratively with such systems is the way to cope with the complexity and data richness of modern life. Thank you.