 Welcome to the AI for Good Global Summit here in Geneva on day one and of course that means we should be speaking to the real specialists about AI and one of them is here right now Professor Mulder from Berliner University who deals with machine learning. What is that? Well in machine learning you teach machines and to predict from data and so for example you could think of having teaching a machine to diagnose cancer as opposed to non-cancer and then you know hopefully this will be a good support for a medical doctor who has this task to fulfill. And what brings you here for this week? First of all I yesterday I organized a workshop with 10 AI specialists from around the globe and we discussed the news advances and this was a very interesting day for the specialists and today I will talk about how to make big neural networks transparent. Basically when we talk about AI for good what do you see in terms of good? Well I mean I work my focus is mainly on machine learning for the sciences so like medical sciences where you help diagnose and where you try to contribute to the mechanisms that underlie cancer that where you study quantum physics in order to find new materials I think that's all in the direction of good. Where are we now in the evolution of AI is this is really just the very very very beginning? Yes you're very right so a lot of people nowadays say yeah and then we apply AI as if there was something you know that you can take from the shelf and just apply it and that's of course complete nonsense. It's a field that is growing that is in infancy and a lot of wonderful breakthroughs are yet to be made. And yet if you look back I mean AI was talked about even in the 40s and 50s. Yes there was actually a long history of classical AI where people thought that using logics and these type of techniques you could actually explain the world and make machines intelligent and it turned out in the last decades that it's much better and much more useful to have statistical learning algorithms and basically they're in all our you know they're in your phone and my phone that's it's basically in our lives. AI though on the one hand are we should be scared of AI or not because it's we're talking about big data we're talking about everything about you everything about me being shared and used. I think we have a pivotal moment in time so I think right now a lot of us are scared right and you know it's it's not that we you know we have to keep things as they are. We're citizens we can think how we want our world to be and I think one of the points in this meeting is to discuss how our world should be and then then there's also the technology people that can think of new algorithms and new mathematics in order to support this. Is AI just like green technology you know competition between Europe the Chinese the Americans to who's can be advanced in new green green technology is AI also something which continents are battling over to have supremacy or is it something that needs to be shared worldwide? I think both are happening at the same time I think the scientific community is through all the centuries has been competing and collaborating and all the great successes were always standing on people's shoulders and the same applies to AI. There's nothing special in this respect but of course AI is very instrumental for the progress of our countries so I think we need to have the people the right minds we have to educate the right minds we have to have the right infrastructure we have to have the correct amount of research funding to contribute to the fundamentals of AI because these fundamentals they are very close to an application it's not like in engineering where you take ten years until something gets into a car after you made an invention it's rather that this takes half a year and even in some companies it takes ten days so so I think it's very agile world but coming back to this you know to the European view I think all of us we would like to have a world where privacy is a respective so we want our data to be respected and I think it's time for having new methods that actually respect the privacy and there's a lot of new legislation to be done and regulation to be done but on the other side there's a lot of new science to be done to make it really practical and because in the end we don't want to kill new enterprises on the contrary we would like to have you know privately respectful AI as a European contribution to the world as professor muda from Berlin University thank you very much for your time I appreciate it okay very good