 Thanks for joining us here in Geneva for the AI Figured Global Summit 2019. And here we are discussing the role of AI to assist humans in making better decisions. And a perfect illustration of that is the IBM Debater Project. And to talk about it, we are joined by Mr Noam Sloanim, distinguished engineer at IBM Research. Thank you so much for joining us. Thank you for having me. And you are obviously the initiator of this project. And this project is all about capturing the narrative of the global population on controversial issues. So why did you start this project? What were you trying to prove or to show? So we started this project in 2011 or actually in 2012. It was in response to call for proposals in IBM Research looking for the next grand challenge. After chess in the 90s and Watson in jeopardy in 2011. I had this notion of developing a machine that will be able to debate humans. It seemed impossible at the beginning. Over time we gained more confidence that this is doable. We worked on that for more than six years and eventually demonstrated the system just a few months ago in San Francisco. Full-life debate between Project Debater and one of the superb human debaters in the history of university debate competitions. And shortly after that we started to develop Project Debater speech by crowd. Where the notion is really to ask what will happen if you will collect a lot of arguments for many different individuals about a controversial topic. Can we generate, can we use the technology to automatically generate a compelling narrative out of that to enhance decision making? So this is what we do. So how do you connect the arguments to data? Because that's an important process isn't it? And it's got to be ethical and to be inclusive and representative of everyone's opinion I imagine. Yes, so I think that people will tend to respond to topics that they care about. So if I pose a controversial topic to you but you don't really care about, chances are you will not bother to contribute arguments about that. So I think this is one part of the question, trying to engage people on topics that they really care about. So we think that Project Debater speech by crowd can allow us to establish better communication between decision makers and the target audience. So for example a government which is considering a policy and the citizens that are going to be affected by this policy. Today really don't have the means to ask them for their opinion in a sophisticated and effective manner. So this is one of the issues that we think we can address. That's a very interesting point. Can we imagine the system replacing votes for instance in the future? I don't think it will replace the decision makers but it will help them. This is my feeling. Because really put yourself in a position of a decision maker that you really want to take a good decision for a change. And you want to hear the opinion of the people that are going to be impacted by your decision. What options do you have today? Not so many. So you can do a survey but in surveys usually people respond to very simple questions. You give them close questions and they vote yes or no then you take the average. This is fine. This is one possibility but we believe we can do way beyond that. So today at the summit you unveiled and it was an exclusive wasn't it? The results of a question you put out to the public in May which was about social media and whether it does more harm than good. And you presented the positive outcomes but it's also possible to find the negative views on the question. Yes this is a good point. So I shared with the audience here in the United Nations one of the two narratives that the system generated. We got more than 1700 arguments from many people all over the globe. And the system automatically identifies whether a particular argument is supporting or contesting the topic about social media. And it generates two narratives. I shared only the one that supports social media. And this was reflecting about 54% of the arguments that we got. But 46% of the arguments that we got according to the system were against social media. So people are welcome to go to the Project Debater speech markup website and see the other narrative. You have been working on this project for a long time haven't you? Around seven years I think you told me earlier. So I started that as a single slide proposal back in 2011. So for me it's more than eight years now. But the actual work started about seven years ago. And it's grown and you're not just working with data scientists and programmers. You also work with people with soft skills because that's important. And I think that's the point we're trying to make at the AI for Good Global Summit. It's focused about hard skills and science and technology. It's also about philosophy, the arts and other areas of expertise as well. Yeah, I think this is a good point. So first of all we have an amazing team of researchers in IBM research that made this happen. So these are experts in architecture of the system and natural language understanding and artificial intelligence and so on. It is true that you need additional people with other skills. So we had people with, I would say, attitude to text. We had even a PhD student from the philosophy department. And it was quite an interdisciplinary team. Quite heterogeneous, which was part of the fun working on this project. One more question to finish with. What's your vision for the Debater? Where do you see it in 10 or 15 years time? I think the Debater is really just the beginning. It touches on a very fundamental problem, understanding the pros and cons for taking better decisions. This is something that we experience all the time. And if you think about it, there are not so many AI technologies designed explicitly to help us take better decisions. So this is what we are trying to accomplish here. I think we made significant progress and we believe that we will make even further progress in the future. No, I'm slow name. Thank you so much. Thank you.