 What I want to talk to you about today is the role rumours on social media play in a crisis situation. In a crisis situation, accurate up-to-date information is absolutely crucial. Not simply to the people who have been caught up in the crisis situation, but those first responder and rescue teams that have to go and rescue people and need the most up-to-date information, having access to the most up-to-date information can save lives. Now, you will have heard the rumour that last year we had riots in the UK. This turned out to be true. But as part of these four-day events, a lot of rumours were spreading on social media, particularly on Twitter. And these rumours included things like animals having been released from the zoo, that looters had managed somehow to push over the London Eye and set it on fire, that looters were frying their own food in McDonalds, which is one of the rumours that has never been confirmed, but also far more serious rumours, like the smashing up of a children's hospital in Birmingham by looters. Now, what I want to say from the outset is that this rumour was false. It didn't happen. But what is very interesting about studying rumours on social media is that it allows us to understand better the life cycle of these rumours. And as an academic, I was involved in a project with a Guardian newspaper that studied seven of these rumours. And what we as academics did is we carefully coded, manually coded these rumours into understanding better whether or not they were simply being spread in the green visualisation, whether they were being countered in the red, whether people were asking questions, is this really happening in yellow in the visualisation, or in grey, where people were commenting? So this rumour started spreading on the third day of the riots, and what you see is in that the first hour of the rumour, it simply gets bigger and bigger, so it takes on quite a lot of force. But what is interesting is that within an hour, you have opposition to this rumour. You have a single person on Twitter using, wait for it, logic, to say it cannot possibly be happening. And I called this the voice of reason tweet. And what this person pointed out is, may I remind clueless, hysterical, Birmingham riots commentators that the Children's Hospital sits face to face with the city's central police station. Now what is interesting for a person like myself who studies social media is the way in which that one tweet managed to cascade through this life cycle of the rumour, and actually took on quite a lot of force, and was able to keep going through the night, and what we see in the visualisation is that the rumour slowly starts to disappear. So it effectively debunks this rumour. So this keeps going through the night, and what we see is that, if we can speed it up slightly, is that it gets smaller and smaller, and some other people get onto Twitter and they start repeating it slightly. But what we see effectively is that around 4.30 in the morning it disappears. People are asleep. People are not on Twitter in the UK at 4.30 in the morning until they start waking up again. And somewhere they've heard that this Children's Hospital is being attacked, and the rumour starts again. But what is interesting is that we see a second cycle of debunking of the rumour. And what we see at around 10.30 in the morning is that the rumour for a second time has been debunked. Now, there's a number of really important lessons to learn from this. Social media is actually really effective itself at debunking these rumours. We see this. We see this now in action. The other thing that's really important to learn from this is that we no longer watch the same channel. We are no longer all glued to CNN. We are connected in different ways, and the ways in which we consume and interact with breaking news is through our own bespoke individual networks. And what is really important as a lesson for the emergency services is to understand the repetitions through which they have to keep debunking these rumours because they have to cascade through multiple networks over a longer period of time, in this case over a 14-hour period of time. Now, earlier work that I've done on crisis communication and rumours shows up the very, very serious problem of those rumours that people want to believe. And what I looked at was the aftermath of Hurricane Katrina, where very serious rumours were spread in the mainstream media that people wanted to believe that had a very disastrous effect on the ground to those victims caught up in the aftermath of the hurricane. So what is very important from the rumour work that I've been doing recently on Twitter is that it's very important to acknowledge the power of individual trusted sources and individuals' trusted networks. And so a journalist like Andy Corvin in the Middle East has built up a trusted network of sources that can help him filter information very effectively. So these things are very important in this current era of big data. I'm making a case for the importance of little data in big data, the importance for human-coded data that actually can tell us what it means. So if I may ask one thing of you, if you find yourself in a crisis situation and you have enough battery life left on your mobile device to send out some tweets, please do three things. If you're going to retweet, check what you're retweeting. If you are going to send out information and you don't know the provenance of the information, please say so. Just say that you don't know, say it's unconfirmed. If on the other hand you're in the possession of valuable information to debunk rumours, add a link, so that other people can cascade this information and debunk rumours. And finally, it is so important to remember that a little bit of logic in a situation this serious can go an extremely long way. Thank you.