okay that was a long break it was a long break and i see some people still mingling and networking and having their conversations so welcome back and thank you for coming back the next guest speaker we have is dr Björn Bringmann he's with partner with deloitte and managing director of the ai institute at deloitte he is also a phd in computer science and over 20 years of experience in artificial intelligence and machine learning so this will be the business perspective on ai and of our situation at the moment and i'm pretty sure bjorn joined me on stage pretty sure that bjorn will share some use cases and implementation cases here with us his speech is called race to ai a global perspective thanks for joining us thanks for the kind introduction um title i looked at the conference and thought uh ai is obviously something we all care about um otherwise we wouldn't be here and this is ray Rise of AI and doesn't it nicely fit to have the race topic in here as
well and it just goes so well with rise to ai and race to ai and i wanted to give a little bit of a global perspective but not only um with some local implications because if you have a race you're typically not alone which means globally uh doesn't really work we have some sections so to speak globally that we want to have a closer look at um and you have seen them before so i will mostly i think augment what we had before i don't really believe that there are many contradictions but let's see so race to ai another analogy that comes to mind and i think that plays very well here is about what's that 70 years ago six years ago there was a race to the moon that was basically kicked started by the us and this one doesn't really feel too different when you dive into it you heard this morning a lot about what implications ai will have on society right um and these are all good and relevant if this
didn't wake you up or if you're still slightly asleep from lunch which by the way was great so thank you whoever selected this they're 12 12 trillion dollar at stake annually just to put this into perspective the gdp from germany is 4 trillion annually so this is globally but it's also three times as high so this is quite quite some money and that's something we're all looking to benefit from so what does that mean 12 trillion obviously this is not just in one niche right this is something that's literally all over the place so this will happen in automotive in banking and anything you can think of plus the things that aren't invented yet at least to say impacts on on social and society now it's usually called a general purpose technology by now which is an amazing term it basically says it's a technology that at the beginning nobody really knows what to do with and then it suddenly is everywhere right such as the steam engine or robots or information technology so all
the computers you have electricity is missing here and they tend to have quite some impact right these are small numbers but if you scale them over everything we have 0.3 is a lot and that was a steam engine and i think you can imagine steam engine kind of one of the first engines had really a lot of impact because basically liberated us partially from manual labor if you look at the pathway ai you can see that it's four times as long right so there's a lot more going to happen than when comparing to the steam engine and another thing that often happens is really this ai will be there somewhere in the future and i have the feeling i'm preaching to the choir here um because i guess we agree that that future is already here there's a nice amendment here which says it's just not evenly distributed we don't have the same ai capabilities and also progress and move an intention behind it globally so if we look at the different players in this
room then you see oh sorry um europe right um so then you see three different players that are usually talked about usually quite many talk about only the two on the left and on the right um quite some though to my surprise positive surprise say and we've heard this this morning and i'm part of that team is that europe is not just relevant but also still has a really good chance to be part of that race or as i prefer to call it have a seat at the table why would we care i think again we've heard this this morning from from hunts and chris and and many of the other speakers this is not just about making something more efficient right this will profoundly change how things are done today and hence it will also pick up some of the parts in this from the society where it's developed in and i might be biased but i do really appreciate human um human values as well that's true but european values right so i
would highly appreciate uh to be part of the game um a question that occurred to me is like race wonderful uh wrong button um when did that actually start and it seems mostly that you know nothing much happened you know there was a lot of ai research and you know all know the story from 1956 where it kind of started as a small group and you know went on and things happened in the background and somewhere there i did my phd nothing really happened and then in 2016 or 2015 rather two things happened roughly at the same time one was that alphago that probably nobody of us played against but all heard about uh won against lisa dole in go a game that everybody thought a computer could never actually play very well and the obama administration also kicked off quite some work in the direction of artificial intelligence which was as it's called a sputnik moment for china because they saw things happening with the game of go probably also quite impressed that this
technology is able to do things that were not just unexpected but really pushed far far away and basically jumped on the rise on the race or started the race in fact and huns i think put it very nicely they have a strategy where ai is basically always on top that's roughly how it looked like when the race started out let's go a little bit more into detail into how that looks like before there's another question maybe where are we racing to right and i stole that from chris who presented that a while ago actually right here and i think it just brings it really to the point these three different regions are not completely um exclusive obviously right but they have a slightly different goal or intention of where they're going to and if you want to put it down to one word in the us it's mostly about efficiency right pure capitalism is if you like um in china it's mostly about social cohesion right you really want to have all as one if
you like and in europe we have that thing called humanism right um so we really really care about the people which by the way is a very nice export model people like europe and come here and don't go to so many other places if they have to leave their home country because here it's about the people so it's not just us liking our values so that's where we're racing too still we're all on the same road at the moment and there's this wonderful quote that comes in um once a new technology rolls over you if you're not part of the steamroller you're part of the road and i really like to sit in the steamroller right and i'm pretty sure uh i'm not the only one and this one really looks like it's just powering up some steam the details are promised not too many numbers right just just a hunch here and there um i looked at three different parts more or less how you get some i ai idea that's probably pre idea
still into production or into the real world where it's producing value not not saying that research is not part of the real world and if you look at those and they came out as three guys having a strong left or a strong right depending on which you look one there's r d spending from the top 2500 computer science information technology industries and the blue one is actually the number of ai publications i think that's from last year or the year before and you can see that the us is by far outspending everybody else right so this is what 10 times almost as much as in europe not germany right so this is all europe um and europe and china are roughly the same if you really want to on on spending the other thing is the right hand if you like a number of publications there europe is somewhere in the middle but there is some addition to make here that developed over the years the u.s publications tend to be let's simply say of
a little bit more value right like however that is measured um european publications usually were a little bit behind maybe and you know amount of value per publication if you like in china back when i did my phd was very much like you know you need a bunch of chinese publications or asian publications in fact to kind of come up with the same value that the publication from a u.s ivy league university would provide that's not really the case anymore they know what they're doing and so do obviously the folks in the u.s and here so they're it's a you could almost say level playing field but the the speed which is changing is interesting here which is um there's a lot more happening in asia than we ever expected this is really good and interesting stuff right if you break it down a bit differently here we see the top 12 i believe companies with the highest r d spending two years ago and you see um well actually let me highlight that you
see it's all tech right everything that's now lit up is tech and you can see europe is in there yay um we're not tech i mean obviously volkswagen is building some technology right but not in that sense and um rush holding is from switzerland obviously and then we have johnson johnson which is u.s again but not tech and somewhere up there uri is chinese and samsung is obviously from korea south korea what's interesting here is two things if you would have looked at this chart seven eight years ago until 2014 then volkswagen was at the top with pretty much exactly the same amount of money they spend now and if you look at the top two amazon and alphabet here they doubled their spending complete compared to the year before so there's a lot of money going into ai research here if we go one step further we're coming into innovation right so how do all these ideas get into reality and that's usually why our startups and that's a very simple comparison i think
the the one that sticks out is obviously in the middle this is basically just a percentage of how many startups are in which region and you can see that the us has i think that's about 65 percent of the total number of startups in those three regions europe comes in as a second not a close second that's i think somewhere about 20 22 and then china is there with 12 percent if you put the others number the other numbers uh together with this then it becomes more interesting so the number of unicorns by the way this this numbers changed while i made the talk so if you double check you will get slightly different numbers um but the differences probably won't change so around about 600 unicorns so one billion or more valued startups in the us almost 200 in china and around about 80 in europe and the amount of funding so that the u.s obviously wins if you like and the amount of funding that goes into those startups china by far outstrips
the other two and that i thought was really impressive right so that being said many things going on and europe is well not really a frontrunner it's kind of hard to say that we're in the middle even right but we're on there the last one on this part is really capturing value like at some point whatever idea you have has to kind of generate some kind of value and we did a study at deloitte last year where we asked close to 3 000 executives around the world in 12 countries which is the state of ai in the enterprise and this is just one of the many chants in there we classified all the companies that replied by the way it was a double-blind survey so nobody knows who we asked and those didn't know that we were asking so as good as it gets we classified all the companies into four different types right which is not really linear but for the stack bar chart that wouldn't work otherwise you have starters that's very simple
they don't really do a lot yet and they haven't really achieved a lot yet right fair enough so those are starters on the other hand on the top the green ones that's what we call transformers they're doing a lot and they're generating a lot of output with this and then you have the two blue ones in the middle the achievers the unfortunate ones they really do a lot or try a lot i i guess i have to say but it doesn't really work or at least they are not there yet and then you have the path seekers that don't do so much but it all works and since we're at the blue ones um let's just highlight two parts if you look at the light blue one in germany that's the underachievers we are really good at underachieving um europe is just about on the same level europe in this case is france uk italy and germany right um so overall europe might be slightly worse at underachieving so to speak but on average you
know it's kind of the same size the other blue bar which is really uh in your face if i may say so is china many that try and succeed right they haven't done everything yet but basically the dark blue ones are the ones that move into the green bar so if you go two years ahead you will see that that green bar in china will have grown massively unless things go wrong and the final piece which i think is interesting here is if you look at the green bars only we're not really that bad either so we do have companies that do things and are successful with it so somebody out there is doing something right since it's double blind i can't tell you whom but they shouldn't be so hard to find right i think we kind of know them so there it does work there are examples to copy but we need to move so what are we waiting for that's pretty much how the race looks like right now you can go
on and read indices of the indices on ai and talent and whatnot but at the end of the day it's always we started roughly at the same spot the u.s was ahead china triggered the race so to speak um and now it's pretty much um a speed car in red the us constantly moving forward and europe's sitting there wondering what they should do so what are we waiting for and i think we heard similar ideas throughout the day but when you look at this race something comes to mind from through the looking glass which is really it will take all the running that we can do to just stay where we are but if we want to advance we need to do a little more right and this is um may sound painful but to be honest working in ai is quite fun so it doesn't have to be is it fear that we have in europe maybe especially in germany that keeps us from being more successful right and they're typically two things right
and um one is the risks and on the other side they're opportunities so things when you do something that will go wrong and things when you do something that will go right which in ai is really risks mostly we've heard it this morning can be over or misused as a risk and on the other side opportunities you know how can you properly use it to generate some value ethics and everything included um and since we're apparently all really good at being afraid of stuff here i thought i put up a third risk which i don't hear too many people talk about there's an opportunity cost which is if we just stay in place we're going to lose something right more prominent at the moment i believe is climate change on this there we have numbers now that say if we don't do something we will use what was it about a 200 trillion u.s dollar over the next i don't know 30 years right it's expensive to just sit and wait it's really expensive and
the sitting and waiting here seems to come out there in in two different forms one is really people like the general population seems to be afraid of it for ai taking over the world ai taking over their jobs part of our survey was on the jobs and to be honest no company really wants to replace any human right because we have and in germany quite dramatically so a demographic shift coming up we'll be pressed to actually find people for the jobs that we do still have and the rest ideally we need to automate to stay up to our level of wealth and comfort that we have here um then there's the other one that i see quite frequently which is in the industries in the companies they really want to be innovative they really want to do something new but they also really want you to show that it has worked at the competitor two years ago already which is not how i thought innovation was defined i thought innovation was defined as let's do
something and take the risk of this going wrong right but let's do this often enough because we know things can work and again we've seen it before there are companies who do this very successfully also in europe so this can be done so i reformulated i mean you may have seen it uh while i was talking here um maybe we shouldn't really fear ai i'm not saying we should all just you know go and do whatever there are many things that can go wrong but maybe it's more about respect right of what this technology can do and obviously handle the the bad things but don't ignore the opportunity cost because that's massive and that's not just massive in terms of us having less money but about uh the different systems in the world becoming more or less powerful and again i like my european values so if we're all fearing it then one thing to at least on population side but also on the on the industry side is really and then we probably need
to do something that we can trust those systems because they're honestly uh not so easy to understand as a knife maybe would be right where we all agree that a knife you can hurt yourself but you don't have to if you use it properly um mitigating risk with a knife is easy mitigating risk with ai up till now is maybe not that easy yet um and there was a huge discussion and a huge uh publication series i can say i guess from the european union on all topics trust with ai and actually obviously as we as deloitte are trying to move the economy and the society forward here as well um we took this by heart and and worked quite a bit on it and came up with these six different areas actually what needs to be considered if you want to build an ai system that's actually trustworthy right that you don't need to be afraid of that you can use that you know what it's doing and so on and so forth so
biased ethics all in there um if you want to know more about this just at the corner is my colleague who's basically working on this quite a bit and i think this is really one part of the game how to move things forward really make people understand that this is nothing to be afraid of another part that would tremendously help is if we would move forward with our digital single market in europe because if you want to scale in the us you have one huge roughly uh homogeneous economy that you can work in china pretty much the same europe not so much there you can see the little black lines of borders right um going through one of those is not that trivial yet so moving forward and making it easier in europe to scale will certainly be part of the puzzle i believe the good news here i guess is you could say once you can make it here you can make it anywhere because once you cross one of those borders any other
border wouldn't really be a problem anymore right so this is a good training ground uh but we can probably make it easier and we we have to to succeed here because certainly none of the countries you can see there will be able to succeed with ai alone another one is um and that's the ugliest slide i made i'm afraid there needs to be more funding for research and innovation i've heard a lot of discussions that the industry needs to come in and the industry says the government has to come in but if you look at the big things that happened in the past be it the internet or the global positioning system that everybody uses all the time now electric vehicles being well they're not really around the corner anymore they're right here right there or mrna which you know most of us got a little injection with recently they were all heavily government-funded because they were simply it was simply not clear enough how to use them from the industry right certainly not by
one player so they're there we need a little bit more cash i guess and just to compare and i thought those numbers were interesting um the the german government put six billion roughly on the table for investing in ai do you have any idea how much we paid for the airport in berlin same number fair enough 5.9 only right but if you look at what it brings i think the airport is probably great um could have been faster uh i think investing in ai is probably a little bit more pressing than the airport right not tearing down the airport now that it stands but if you compare those numbers not so great um then we have this again two for spun in watson which i believe has about a billion for the next five years give or take harvard in the us has about two billion per year so yeah a little bit more probably won't hurt and another thing um again i said this before can't do this alone and i also don't believe
that one company can do this alone and there's a famous example obviously of silicon valley um were there many things happening at the same time right probably start with xerox park years back and then other industry settled down there you can go into a bar and people from facebook will talk to people from google and so on um many startups are there there's ideas floating money is floating and all this kind of stuff and these companies probably wouldn't have done it alone so having an ecosystem sounds like a really really great idea and again i think it's not just necessary but also can be a lot of fun and the different players we will need in there are you know the ones mentioned before we will need society in there um we will need the industry in there because quite frankly they have the money and they they need the disruption we need the research in there because well they know what they're doing simply speaking and the government because at the end of the
day um that's all of us right and those being connected and here's a little um little plug um there's a summerfest from the kai park on friday so if you don't have anything planned or put differently if you have something planned cancel that and go there that's a growing one of the growing ecosystems here and and it's going to be great one thing that i think is left is probably the most important and has been said before but can't be said often enough this is not going to work if people don't want it i do fully understand that people are a little bit spooked maybe about ai the first algorithm i built 20 years ago spoke to me quite a bit but people don't have to be afraid of it because it's not really doing anything badly we obviously have to be aware of it and respect it and make sure that nothing bad happens in the future but right now i think one of the most pressing things really get people on board enable
them to understand what we have there and enable them to use it and to push it because again if the society doesn't want it it's not going to happen and it worked before if you look at this picture i think it's fascinating um two guys obviously kind of outdated fashion by now um this was 1945 when 1954 sorry when cern was built and cern is the place to go if you're a physicist why wouldn't we have this for ai thank you thank you girl we have some minutes left and um if you have some questions or also in the live stream happy birth will probably be happy to answer some of those any questions opposing opinions we heard a lot about again about the spread between europe asia and u.s and we have one question coming up here i only followed the last part of the presentation the part on science or technology and society interested me the most with a nice picture of smiling faces how practically would you do it and could you
give us some best practices which are the governments or the institutions who do it well communicate with the society of midburger thank you yeah very good question so how to do that how to excite people um i think we've seen uh one presentation earlier which was about is ai evil um i think you get a little bit of the drift you know all the termites terminators in the movies uh don't usually awaken a sense of i want that in in people and i think we're going to hear a little bit more about how this worked in in the nordics um what i'm aware of is i think helsinki has set up a course for the public to it's free you know you just sign up you go on the internet as you do sign up and you go through this training and it's tailored for non-tech people it's tailored for the average you know joe and they aimed at getting one percent through this so one percent of the finnish population was the hope would
take this course to really understand this a little bit better they reached two percent by now now two percent doesn't really sound like a big number but they overachieved by a hundred percent what they had aimed for and this course is now available i think in any language so most simple thing go there sign up i think then after this it probably should go a little bit broader into communication as a whole right so not just you know me saying sign up to that website but really bringing it into companies that's what we see quite a bit that in companies here obviously have a push back do all kinds of fears if you talk to people first what it is about um they will get excited maybe not all of them right but you don't need 100 if you get 99 that's fair um and again i think if you if you educate people more it doesn't have to be the super complicated technical stuff right um you don't need to know matrix multiplication to
be not afraid of ai but you can bring this in in schools universities as a course doesn't have to be you know full-blown ai foundations course so classical to be honest there might be better methods but if we start with that it's not a bad start thank you another one we have time for one more question thanks vern for the great perspective and i especially want to like ask again about this like opportunity cost and turn this a little bit around and say what do you think does it take to make our large industrial corporations get a fear of missing out showing them the slide um i think they already do right they might just not be afraid of it enough but if you look at the car industry there's one automotive manufacturer that's basically a tech company and not an automotive company that's scaring people and they build a factory right around the corner um mr dies apparently has taken up the challenge and that's good i think they're also the ones being the
only ones on the on the r d investments right how to turn them all around i don't know i do believe some will probably just disappear um and if that's what's happening and it's not happening to all that's probably fine and i think we've heard it this morning also from chris who said yes the banks push back and say we have sensitive data we're so over regulated but then google again also has sensitive data and is quite regulated and they're just thousand times more efficient if i would be a bank i would probably listen to that sentence more than once but there is there is a real threat have i calculated all the use cases for all the industries no but to be honest i don't think you would need to if you listen to these numbers so sorry for not having a easy how-to but complex topic yeah it's a super complex topic it's a super complex topic yes and bjorn thank you so much for starting us off with the topic of ecosystems
we have the panel coming up now thank you so much and was a great business perspective on ai thank you