 So I'm sure some of you are familiar with devices such as Alexa Google Homes. I'm sure some of you have them, wonderful listening devices at home. It's kind of been interesting to see the growth of these devices over time. Actually, there was a kind of a recent bug in the Alexa, which I just found hilarious where it randomly laughed. But I'm sure you could kind of discuss that. But I'm going to go hand it off to Mario to moderate the panel on kind of the intersection of open source and the conversational technology. So feel free to see it. Thank you very much, Chris. Appreciate it. Okay, so the conversational web, that's our topic. And of course, we want that web to be open. What is the conversational web? We tried to figure this out here in the panel and a few more questions. So, yeah, I'm the panel moderator here. I'm also happened to be one of the first Asia founders. And our panelists are Michael Christen, Michael Christen, founder of Susie AI. Please come on stage. So Michael is the founder of Susie AI, but also a lot of other projects like Lokrak, he's the creator of the peer to peer search engine Yassi. He's a big data engineer consulting for some of the largest corporate players in Germany on search technology and digital transformation strategies. He's also the architect of large search portals like the German Digital Library. Welcome Michael, please sit down. Then we have Ken Friedl, data scientist at Daimler. Please join us. Ken is shaping the field of product data management using analytics and AI methods. He was born to a German-Japanese family, raised in Germany and spent some time working for several enterprises and research facilities in Japan also. Before working as a research associate in robotics at the Technical University of Munich. So thank you very much for joining us. And last but not least, Rasmi Ranjan Mohapatra. Please join us. Lead technologist at OCBC. It's a bank here famous in Asia in Singapore. And your primary background is in data analytics, system design and product management. Secondary background covers engineering, quality assurance, user experience, infrastructure operations, information security, and customer support functions in both corporate and startup environment. You're currently leading the technology stack at the open vault. It is OCBC's banks, FinTech and Innovation Group. So you have listed a lot of interesting subjects in this area. And like interest in corporate innovation, applied machine learning, blockchain, of course as a bank, right? And yeah, your current interests are with chatbots. And actually that's very interesting for us then as we want different people here on the panel. So thank you and welcome. And we go back to the introduction slide. So actually it's a very interesting group that comes together here because with different backgrounds here. And what we see with open source, it went to a lot of areas like back ends. And yeah, we have open source browsers. We have open source many different areas on technology. So the question is now with the conversational web. How I would say the conversational web, what is it? Well, we had the graphic user interface web, say traditional browsers, but we see people are more and more talking to devices. So it's kind of an evolution, whereas the graphic user interface, it will not go away, but we have the conversational interface on top or maybe for certain use cases. So what's happening in this area and how do you see the possibilities of open source? These are some of the questions here. But maybe we go around first a little bit like maybe you talk us, you tell us a bit about your personal interest of like chatbots and like AI applications in your area and what is your personal interest. Yeah, yeah, Ramsi, Rasmi, thank you very much for correcting. Actually, like today also somebody said to me like, oh, everyone said the German names wrong on the first day. So my interest in AI comes from my schooling background. So I was a PhD student doing computational biology. And I started as a looking into a gene regulatory networks, how the time series data of genes are behaving for a time. And then we are correlating them to produce some proteins and molecules that in the future can be correlated to a drug. So that's where I started. Then I moved into industry, but industry is a different kind of environment altogether. So the first experience, let's come to the point, conversational web. And has always been seeing the web interfaces in terms of a GUI always, right? You type in, you get a search result. That's where Yahoo started out, became a big hit. Then Google came and simplified nothing white. Then you get a top rank, the top 10, top 10 relevant or top five that you are interested in. Now the modern generation is even bored with that. They don't want the top five to be there. They just want the first keywords that is relevant to me. So I see it as a gradual evolution of the way people want to interact with the systems that they have, including mobile. And as the way you started it out, basically not all the use cases will be conversational. Nobody wants to tell that my account number, please trust with this amount because people will be hearing it. But there are use cases, for example, Alexa plays 45. Nobody wants to go, click, and then do it. So this kind of conversational interfaces will be more common and common. There was just a session about cars and how to automate them right. So it's already there. And amount of use cases that is going to be in future, it's tremendous. That's what my personal belief is in. Ken. Hello. Conversational UIs. So you're from Daimler, you're working as a researcher a lot, but you also connect with other teams at Daimler. So you're in a very good position, you have an overview of many projects that are going on. What's your take on Conversational Web? I know Daimler released M-Bucks in English. Maybe you can tell us more about what's going on. So I'm basically working in the field of product data management at Daimler. So first of all, what is product data management? It doesn't sound very sexy, but it's basically connecting all the data that is generated in the process of developing a car and also in the process of producing and later on using the car. That's the important part where it comes to the interface with the customer. So we are trying to manage this data in terms of a digital twin. So you have a digital copy of the car from the beginning where it's manufactured to the end where it's basically, well, destroyed. And of course in the middle where it's used by the customer. So my first interest in conversational and chatbots came actually personally because I had a... Actually at home I have a small R2-D2 robot, like 40 centimeters tall. And I wanted it not to only run around and do SLAM, which I had originally on it, but also to talk to me a little bit. And I was motivated by the story of Eliza and then I thought, okay, I can also do that maybe. So originally I had a TensorFlow chatbot on that and then just talking about the weather. And then I abandoned it because then I got a new job. But anyway, so in terms of Daimler we're interested... As far as my understanding goes, there's like three different types of chatbots. So first is the service chatbot. So you can ask questions about some sort of manual. In terms of Daimler we have Ask Mercedes, for example. The app that was launched end of last year in several markets, Malaysia, Indonesia, South Africa. And so you have an AR overlay over the cockpit. And then you can see if you have not been a Mercedes driver before. For example, how do you charge your phone or how do you activate Vipers or anything. In my case, in my professional environment at Daimler, we are interested in the second type, which I would call the process-oriented chatbot. Where you're trying to teach, where the chatbot is basically a teacher, teaching a person how to deal with a certain program. I mean, these processes can be very simple. Like just get a vacation permit, start and date, and that's it. Or they can be very complicated. Processes such as developing a new part in a CAD environment. And personally again, coming back to that, I'm of course interested in the third chatbot that I think is an entertainment chatbot. You can talk about whatever you like to a device or in my case to my R2D2. Actually, originally I had even, there's this R2D2 translator online. I had converted the text output just for fun into R2D2 text output, but I couldn't understand it. So like, beeping and something. But of course, there's another topic for translations for chatbots. And then I had a male voice on it, and then a female one. It all sounds awkward. Actually, the original voice would be the beepings. But anyway, okay, humans can't really, and I can't really learn that language anymore. So yeah, getting back to the original voice. Okay, thank you. Thank you very much. But you call that chatbots. And maybe that's something that Michel also has a take on. Is it like chatbot? Is it something more a framework? Or you are working on the conversational web with Susie AI. What are you doing there, and what is it? How does it come into play here? I was always fascinated by thinking machines. And when I was a student of computer science, I concentrated on logic and deduction of automated proving systems and programming in Prolog. So this is now a nice coincidence that after many years when I did search engine technology, I was with Yassi and Lockluck, I was thinking about an additional layer where we can have a logic processing of search results. So Susie is now the latest baby of open source projects, which I'm working on. And it's actually the realization of this idea that having a search engine aggregation service with a logic layer thinking about results. And it's not like a chatbot, it's like an expert systems on information in the web. And it's a really good coincidence that now the compositional web is exactly in that field where we can create tools in that field, creating expert systems about topics people want to talk about, getting information from the web and so on. So this is a really interesting field to work on with a kind of thing that's coming truly from artificial intelligence theory. So the compositional web as a replacement of user interfaces with just the opportunity to talk is an amazing field where we don't know where it's going, because it's like in the beginning of user interfaces where we didn't have a standard way to communicate with a program where there was the invention of a dialogue. So it's not a standard way to click through a status to reach a specific goal that the machine is doing what you want to. But it's a dialogue system so you can... it's like speaking with the machine with a graphical interface. And now we are speaking with the machine with natural language. So there are no standards. So maybe we could create a foundation of what a standard could be doing this. So Ramji, at the introduction of Chris here, he mentioned like Alexa, Google Assistant and so Cortana Siri, they are the solutions of these established players and they are trying to all push their own. What is your organization already using any of these technologies? Are you working with startups or with OCBC directly? Do they already have apps in this area? So this is a pretty interesting topic for financial institutions, right? Because if you talk about any banking or financial institution as such, data for them is very, very critical. It comes under regulatory purview. You lose one data, you have to get a fine, your reputation is at risk. So banks are looking at these chat boards as a retail point of view. When I say retail, it's basically day-to-day customers doing their internet transactions. The chat boards or the conversational web as we are talking about, this is no way, in our point of view, is going to touch upon core banking system. And when I say core banking system, it's like huge, huge machineries that is actually running the bank. So it's pretty much on the front. Now talking about what OCBC is doing, so we started up FinTech and Innovation Group, the group I present just two years back. And chat board in the first year itself, it has been a pretty high point for us in terms of KPI to develop that capability inside. So from the day one I was involved in the way to have that understanding how to build a chat board, obviously the answer was Google, Facebook, they want to do it because they are data collectors. Now you are conversing, you have no idea whether Alexa is listening to everything that you talk in your home, or it is basically listening to exact question. But that's the business model. So for us, it's more critical in the way that we listen to what means to us. So we are not going into hardware development, like developing smart speaker and then make it like OCBC speaker. And that's your banking friends henceforth, not like that. But what we are doing, yes, it will be more on-web, it will be like a web pop-up where you can query very specific questions. For example, we work with startups because it's very difficult internally bringing in this. So the whole world, the banking sphere, they are collaborating with the startups to get all this knowledge. So we work with the startup, which is based in Spain, and we created something called MI. If anybody is an OCBC customer and they have recently applied for a SDB or home loan, they might have gone through this process. So it's basically a web interface. You talk to Emma, and Emma will be guiding you what are the loan application procedures, collect very minimal information that doesn't valid any regulatory norms, name, address, not your account information, et cetera, et cetera. It directly links you up to the Relationship Manager database. Next day Relationship Manager comes. The information that we collected about the customer is updated, and a call can be directly made. So what's happening now is in one and a half years before, a customer, if he wants a loan, he will actually call OCBC, and OCBC will be a Signing Relationship Manager. Then you start your conversation. But now the whole chunk has gone away. The returns are pretty high. We have already gathered more than 100 million from this channel. The user and customer activities that I'm talking about. Apart from that, there are other chatbots that we are working on, but I would like to keep it confidential for now. It's cooking up. Yes, but we are very interested. Of course, this is a tech conference also, in the technology stack that you are using. What kind of chatbot is it? How much AI are you already applying? The definition of AI changes over time, so there are much higher expectations nowadays. That would be interesting. Could you give us some insights here? Is it rule-based, or how far did you go already implementing these services? I will classify it in the category of semi-supervised kind of environment, because the way we are not going into deep learning modules, the customer can query anything, and we have all the information to pump into that chatbot. In my opinion, the only guy who can do it very beautifully will be Google. The reason for that being, is that they are calling each and every information, be it in white or black, they are calling each and every information. So they have the databases of that in a structured tree or graph, whatever you want to call it. They can go into it to the granularity level of treaties, zero level, and then pull up that information. But for us, investing those amount of money to showcase our retail customers that we can do the business, doesn't make sense. It will be the same for all banking and financial institutions as well. Of course, the industries we are using. At certain point of time, there may be a technology which will be able to just put that SDK in your platform. It crawls through your entire database as possible, makes a giant graph of it. Now you query at any point in the end of this graph and it goes to the other end of the graph and puts that information. That will be awesome. I'm looking forward to it. But probably we are more than 10 years or 20 years away from having that kind of neural networks who can fetch anything. So the perspective that we are going into is make it very specific and make it semi-supervised. So what I mean by that, define the problem statement very, very well. And that means in technical words, define the keywords. So in the world of NLP, the synonyms and acronyms and the way a semantics of a word is developed in the data dictionary is very, very important. If you are not accurate enough to develop that, you are literally not able to answer the questions. It's like the conversation will be broken. You go one by one and then, sorry, I can't answer you. And then you close the chat board. Because nobody is interested to have even one wrong question or not a perfect question. But the chat board replies like, sorry, I'm unable to understand. Me personally, I close it because I think the chat board is not smart. I think people also do that. So in order to have that very concise to-the-point answer the strategy that we are looking is make it semi-supervised and define the problem statements, keyword, practice it in forms of a pen and paper drawing board first. The way you will ask questions, though we are talking about AI but nothing beats the background of the domain when it comes to building such enterprise-level applications. Yeah, you can have fun going through it at home. There may be some answers. But if you are actually serving customers through this you better make sure that it is 90% accurate. Otherwise nobody is going to use it. It's lost interest. No, I also heard many people said actually below 90% is not usable. And yeah, I find it very interesting. So the security and privacy topic, that's one reason that you can't just send audio data to Google. I guess that's obvious to many people and this is a question for many companies here. So, Micha, is that one thing you're trying to develop? You say you come from search engine technology. You're developing the Suzy AI framework. Is that what you want to cater for? Companies like OCBC, the beautiful SDK that you just run and they will deploy it and use it and it just works? Yeah, if you say Google is the only company doing this would you use the Google service into your own company to solve the problems of your customers? Because it gives too much private information from the customers to Google and you would first have to ask your customers do you want that all the information you put in is sent to Google. So obviously Google is able to do this but it's not the first choice if you want to implement that service. But actually you're true. Google has so much information that all these big problems can be solved there. But as I'm working in the open source area for a long time my approach is that if this task is important enough then the community wants to do the hard work. So if it's about collecting synonyms I hope that the community catches up this work and is doing that work. So I don't know if it works. But what's always important in open source development is that it's truly a thing that you can give people and they implement it in their machines, in their homes and maybe send it to Mars because they don't have a cloud access to Google. So being independent is one of the key elements of every open source element. So I try to implement this in Susie as a thing where people can extend the elements of it. But it's still hard work and big companies are in a better position to solve these big problems. What insights can you give us from Daimler? Like Ramzi also said, some parts are confidential. It's not ready for announcements. But you released different conversational interfaces recently at Daimler. You're also developing your own system. So the question is how does an enterprise work, like Daimler work together with open source software? The question is like MBUX for example. I couldn't get so much information online actually what's the technology behind and how you develop it or who are your partners in this project. So how does it work? Why don't you just take Alexa and use it and say okay, we implement an Alexa skill. It sounds like okay, you have a very good skill. Why don't you just do that? So I think generally speaking for an enterprise you need to know what is your core that you want to protect. If you use Alexa services or Google services, commercial services of companies which are largely interested in collecting data, then you need to ask yourself if that data for the company is of a very high value that you really want to rather keep it. And then work on it by yourself. And then solutions like Zuzia would help here of course because I mean so it's not the core technology of Daimler to develop chatbots. But these chatbots would definitely collect data that would be very, very important to the company and that could be worked on later on. So I think it's a... As long as the... I think open source helps in this way because you can really define the technology to have the conversation and to basically have an API between what's going on internally and what's externally. And it's a better way to tackle those things than to use software that is produced by companies which are mainly interested in collecting data. So we are here at FOSSAsia, so free and open source software. So of course we always want to know what will be the future, what will be the answer, how will the open source community prevail in more and more areas. And we've seen a lot of successes but also we've heard from some people that in some areas things are moving back. Like some people said like now it's hosted on a few cloud services. And so the question is will the conversational web be open or not? Well, I can tell from a perspective of a user as well. Okay, so I have one device and now I use this device at home and this device knows my preferences and knows the music that I like and I don't have to have a long sentence. Sometimes I shorten the sentence it still understands what I want because it learned. Like it understood me better by my pronunciation or whatever. Now I go in the car and there's another device. It doesn't know my preferences. I say continue playing my song. Can't play my song. It's not connected with the other device. It's a separate service. Now I go to the banking website and say I've already authenticated three times today. Now the banks tells me, okay, type in this number. I can't just say, okay, I authenticate with a keyword and with my voice. I mean depending on the level of security you can actually use more and more ways to authenticate yourself. So voice could be one additional one. And yeah, so these are the questions. I mean even like not open source or different proprietary ones, aren't we again looking at a place where just a few players will dominate the market and you have to participate just with the browsers. Many years we had to support ActiveX. Banks had to support it. And could this happen again? Will this happen again with a conversational web? What's your take on this? Well my short answer is yes. It's going to be dominated by big players. But that's how the reality of the world is. And the reason for that, because the way you are explaining it's like a lifestyle adoption. You wake up, your wake up is, let's say Alexa wakes you up and then you go to kitchen, Alexa tells you how to squeeze a lemon. Or you tell it. Or how to boil the milk and then you drink and then you come to your car. The same ecosystem, and what I mean by ecosystem, it means essentially you have to tap onto the same players all throughout your life. And these are core business for these people. If you talk about Alexa, it's Amazon big giant. If you talk about Google, again big giant. Facebook, again big giant. So whether a car company or a bank will prevail through this, the answer will be very hard to say now, now. The reason for that being these are companies which are not in the same core businesses. They are competing, they're not competing with each other. And hence there will not be a single entity or a government or regulator telling that henceforth you guys do this. Well, the situation is changing. Like in Europe, GDPR, the new regulatory norm has come up, the so-called explainable AI. So what does it mean? It means, let's say you are applying for a credit card and your credit card got rejected. So now the customer can query to the bank and the bank is regulatory liable to answer all the points why the credit card was rejected. Similarly, in future, when all this conversational wave is coming in, I'm just going in back to one step back. Zuckerberg was apologizing yesterday, right? Because he collected some data. He shouldn't be. So now what's happening? If you are listening to each and everything of the person's life and you want to have that big moment of time that usability is questioned, I don't think so. It will be very, very hard to do. And hence there will be very discrete players who are good in that particular segment of industry that's going to happen. Thank you. Can you look very contemplative? Right? Okay. Yes. What do you think? So I think, so I mean we are just at the start, I think with chatbots. I mean also for Mercedes, ask Mercedes one of the first things, many more things to come. We are also at the start at the new way of thinking coming from a company that is, the main product is the car or the van or the truck to a company that has also is more service oriented and software based. So and one important issue here is of course the data because like the question is again, is that the data is that representing a core part or a core interest of an enterprise because data is money in the end, right? And I mean, okay I personally think there should be some regulatory instance here which or some agreement between enterprises to which extent can the data be shared or I mean there should be of course in terms of usability the best would be all the data is at some central place, right? But then the question is who does that belong to and or is it, you basically have a big data lake or a bunch of companies throw something in and by the percentage of data you can say how much revenue is getting from that. But here we are again at a very early stage so it's very difficult to say I think I don't know what will really happen in the future but definitely in terms of usability there should be something like this sharing the data among enterprises and also defining very clearly for companies who are not originally from the software from the IT part that what is the core in this field. So we also hear that like sales argument for enterprises and for customers is if the partner or the seller stores less data, yeah? So even like I'm as a customer say okay I go to this company they store less data for mine but like we need this data so it's a funny thing we need this data to make all these services work however like the question is which data really has to be stored like let's say on the server of the company or could actually clients store the data? I mean not the data many different forms of data but like are there architectures could we think of architectures technologically that store like data on the client for example like let's say you have you also mentioned another advantages open source you have the code so you could actually say are they really doing what they're saying? So that's open source but like could we say for example like some personal data like what song was played last on another channel let's say is stored on my client and I'm going to the car and the car maybe uses another infrastructure or another company but it can't say okay this was the last song or something like that there's a big problem with the usage of a large data set or data in a data lake you have in that area where we are talking about it's all about personalized data and there's a new law in Germany saying that you're only allowed to store data if it's related to a specific service you're offering if you're not offering the service using that data you're not allowed to store that data anymore and as far as I see the approach of companies is now not to delete the data but to offer new services which are using all these data elements so it's a so there's the need from many companies to create new services because then they are allowed to store data and I would like to catch up an idea you had about what happens if you go to your personal assistant in your home then you go to your car and then your bank and this would mean that the personal assistant moves its state information from one instance to another at either we need a standard to move a kind of mind state of your personal assistant and furthermore maybe you don't need to go to the bank anymore because you can send your personal assistant to the bank and you need to hand over all your keys to the personal assistant as well so this is there are many open questions like how can we create standards there and how can we create security and so that's a difficult field and another point I want to mention is that personal assistants are gateways to markets so if you are in a home town if you create a pizza if you have a pizza restaurant and you're sending pizza to customers then if one of the customer has an Alexa device and says Alexa please send me a pizza then it's not necessary anymore that you are creating a very good pizza but it's necessary also because I know that you are there so the other way around the personal assistant providers are now choosing what kind of products are in the market and are available to the customers and then they use your personal preferences and so this goes all together into a big problem how can we make this interoperable all together without using privacy for other people so interesting ideas and their standardization committees and we see examples in other areas but my question would be how could let's say an independent body how could that be funded so apparently the big ones as you say they are at the beginning things are still at the beginning so I think it seems to me everyone wants to push through their solution they also don't care about interoperability so you create a skill on one channel it doesn't work on the other so I don't know so that was the strengths of open source when I look at the browser wars in the end like openness prevailed so how would that work out here I mean like we had Netscape and they got like this big funding and started like Firefox or Phoenix at the beginning so in regards to the open conversational web if we actually want to pursue this kind of idea how would we do that it seems to me like it costs a lot of money standardization committees they have to meet they have to get together they have to talk to companies it's a process Mozilla was able to do that but yes Mozilla was about breaking a standard all the browser providers actually created new HTML tags so people would choose to use these tags and then that browser was the only one to show it in the correct way but here we don't have any standard at all everybody is doing their own thing so we need a standard for the conversational web how to have a user interface standardized I think when we are talking about conversational web the last word is still web right so we haven't we haven't really gone out of that browser yet yet and that conversation on what is happening it's still a lot on the web interfaces right and again to backtrack on the standardization it's always about the data the current winners in the browser be it chrome internet explorer incorporates or whatever it is it's still revolving around the idea of marketplaces it's still revolving how you can create an app for developers how you can provide an integrated IDE that is web interface so you are providing more channels through your web interfaces even Google created the chrome ecosystem chrome laptops it's on web it's all about data at the end of the day right so again there is no standardization there itself the more these guys are getting of course the more and more services will be coming from them I think it's only since last couple of years that governments particularly in Europe are started looking at it and they have started finding Google was in deep trouble of 12 billion dollars just one year back in Europe right and lots of legal questions government is now looking at the valuation of data itself so once the initial standard of how computing itself is happening that comes through then it will be all tangled up into a more coherent holistic point of view and that will define all use cases wise because standardization cannot be generated if you make a generic standard better make it people will have ways to go through it and create their own standard that's where the beauty of open source comes in so it's very difficult to understand how this ecosystem will play unless maybe if if in this panel there was a government body somebody like in Singapore there is IMDA from media development agency I guess so these guys basically are the owner of the goptic of data so this probably they are looking upon valuation of how this government rules on policies are changing around the ecosystem of this big players so the answer very complicated no idea so if we think about like business models that would like encourage open source or let's say like an open source business model that's not based on data how would that look like is that actually possible or because if it's not possible what I see in like projects that succeeded mainly they have like different bodies they have community of course but like they also have different like sometimes one company sometimes different companies for example the Debian community like a lot of companies offer services like it's a very healthy ecosystem but if we don't see any like commercial way because open source people tend to like say okay let's collect less data but if that's where the money is so how open can we be like what open stage can we reach and how successful can open source be yes can okay why not I mean okay there's two questions here one is the privacy and also can we go to I mean generally we should ask ourselves like what is what do how is the relationship between the user and his or her data I mean we should accept that data is basically like money you know you can spend it and then someone okay and in case of data also once someone has it it has also the power you give a certain part of your privacy in terms of power to another body but you also get something from it and that's the convenience and that's the ability and your car you I don't know you enter your car you don't even have to tell the car where to go it already knows so this debate and and the other question is about yeah I mean where is this all leading I mean what what do we want the chatbot ultimately to be you know do you want it to be just a someone who is like a co-pilot or an assistant you want it to be another social entity you can always entertain yourself with or yeah my good question you told me about the film her in the past like you're a fan of Star Trek is that what you want or what's your idea what do you want it to be that's actually true that science fiction movies and movies are giving us the ideas because they had been creative minds which create personality in movies where they think this is a useful this is a useful entity in this movie so movies gives gives us a good example how these kind of these kind of artificial entities could work and there are many completely different examples like like hell in 2001 which is like another person in this ship it's evil but it's working very well in its way and R2D2 is another example it's the computer on Starship Enterprise is they are all very well examples of where can lead to it's a bit funny but it's not a bad idea to look at these kind of movies and her is also a very good example and it's the example of her is not so far away if you take out the love story then having a conversational interface as an operation system which works on your emails and your personal data such as the web for you some organization working flights and this is a very good example okay great so we have a few minutes left here for the end and I think we can open it up to questions also from the audience or to any statements how open will the conversational web be what does the audience think or do you have questions to the panelists no questions because I see no okay so there we go it's great to think about what we can do with technology but given the justifiable concern of populations and governments for data transfer how do we see this playing out in a fearful world once again technology comes to rescue so I think the technology name is homomorphic encryption so what it comes through is the world has always been fearful how do we share data with each other because data is money it's your core assets so there are new technology players nowadays who are defining this in this way for example let's say A in terms of binary is some string and then you encrypt it the string becomes something else then there is B let me simplify let's say 2 plus 3 is equal to 5 so 2 stands in binary 0 1 and then 3 if you encrypted the homomorphic encryption the result will be computed upon encrypted data but not on exact value of 0 1 right so I have personally met recently a couple of very interesting startups who are trying to crack this again there are also bigger players who are not only coming through this fully homomorphic encryption technology but they are also bringing in the legal team along with them there is a startup I met they call them some technology startup but 21 team are lawyers 3 are developers and rest are CEO and CXL so you can now imagine the amount of effort these guys are making to bring this cross-transfer of data into play but again people are protected about that data those who are senior managers it's very difficult to tell them that it's a technology game and you can win it's not it's very hard hard world so at some point of time government has to jump in and say as Michael was telling the standardization and once that standardization comes with this technology etc it will be prevalent privacy is a big issue for every company and there should be a very high moral level of taking care of the privacy of the data of the customer so if you don't see that if a company has not that moral standard and gives away private data against the will of the customer and that company doesn't break afterwards then we have a very bad example that privacy is not really taken care of and that every company can do whatever they want to and so we will finally lose that that privacy war against what we actually want so this should be a high standard and we should all take care about it the same must apply to governments so if governments don't take care about privacy it's a bad government yeah that's true I totally agree at least at Daimler we take this very seriously I'm not just saying this to make commercial but it is a very serious issue of data privacy and where we want to go in future and I think this is definitely a topic you can't control everything in this world and there is also a level at some point it's like in every relationship you need to trust but definitely enterprise and also governments should really take care of this in a very standardized way and the more data is out unfortunately the more governments could also use it to suppress that's true another question thanks for the talk you had there and always around the data issue would it make sense to show the consumer the value of his data in its ongoing transaction to make him aware of what is really going on to have complete transparency transparency is always good because it creates trust maybe not just transparency it's real value in money but it's difficult to say I think it's true you should maybe there should be the customer should understand what the value of the data is it's like also within a company we share information with a certain other section we know that's helping that other section and here I think that should also apply for the customer enterprise relationship but it's difficult to do as Michael said so how do you measure it in money and then the question is from the customer why do I not get the money and why it's a company or like why don't I just not give them any data and just pay that amount of money for the same service it's difficult but then again the service can't be done without the data never mind my last point I think it depends on how educated the customer is of course there will be a tiny percentage of customers who will be asking this educate me and tell me why what are the decisions that you are able to make with my data and organizations are really really protected about this and they will be because it's great and better for them so the only way you will make it happen again government has to say that okay you have to release you are making money out of this guy if you ask how did you make money explain to him and that's what is happening in Europe as well GDPR policy changes now the customer has the right to ask and to get a full review of all the data that the company has that's true so I see we're having a topic for a whole new panel here about governments and I think like we haven't even touched this if we look like also to countries like China where it's a completely different story also for conversational assistance for governments and data privacy so I think one thing that we see in the open source community people trust in the code so if they see the code well not everyone can read it but like to know that it is possible already is good so let's see what comes the next panel I think we're opening this topic and we haven't even like touched a lot of the questions and see what's coming up what will happen so the future will be interesting thanks a lot for joining here and please give a round of applause to our panelists so we will continue here in a few minutes