so next joining us is a researcher al joshua borchardt i have to read this he is the deputy side director for the german research center for artificial intelligence here in berlin and he is an expert in language and artificial intelligence so perfect uh host uh for for this discussion and the second one is mario brandenburg he's a member of the german parliament the deutsche bundestag from the um free liberal party democratic party so of fdp um and i also have to read this he is this policy spokesman for the parliamentary group of the three democrats in the german bonus talk and a member of the committee of education research and technology assessment as well as spokesman to the committee of digital agenda and study commission for artificial intelligence so those two um we'll discuss today this very interesting topic if germany can actually make a.i so are we just a paper tiger as i think the literal translation is papiertiga aiming for the moon or are we actually capable of doing this so this is
a bit of a controversy maybe if you have questions as the audience please use the chat and ask those questions i can then later on give them to our two guests but now i would like to welcome alyosha and mario and please cheer at home and clap for them for our discussion hello hi how are you doing fine also you want to start basically go ahead yeah we are fine so we first of all we didn't know whether we will have a moderator or whether we would just be talking to each other um so we said it's a little bit of a surprise but i mean i mean we could start maybe with a home story and and tell how we got to know each other mario and me that makes my job easier else i have some questions for you whatever you like yeah you have some questions but but let me go ahead because it was it's it's fun i guess so i i think two and a half or even three years
ago dfki the former head of dfk ibov gangwadster got an invitation from christian lindner head of the german liberal party uh to join a brownback lunch and discuss about this new hot topic this trend of ai right and then but unfortunately wifken waltzer could make the appointment for some uh scheduling conflicts and then so he went to me and he said how do you do you want to talk to christian lindner about ai and i said um yes yeah okay of course if you think i can do it then i also think i can do it so and and in this room was mario and it was a real a real nice atmosphere christian linda in the very beginning says i'm the boss but the other people in this room are the experts for example mario so whenever you want to talk about content talk to them when you need political support talk to me like this i mean so he was very friendly and inviting so we talked about ai and after this meeting
as always we say oh yes let's stay in touch and so on and so forth and then we we departed and i said yeah let's let's stay in touch well we will hope no not hopefully we will probably never meet again i mean like like this always happens but then half a year later this um on cat commission of the german parliament was was founded and then mario and i i don't know mario do you remember how this it went on then did you ask me did i ask you i i actually asked you because alyosha was the only one in the room who um basically was reading the live 3.0 from max tech mug and i guess we were reading we're reading it to the same time and he was the only one as well being a bit of a dreamer and fascinating around what could be and can i fly to the moon and find new planets and blah blah so basically we got to talk about that topic and when i was
um in charge to nominate somebody for the commission um i remembered joshua and i was fighting very hard to get him in because i think it's important and maybe that's um that's that's a transfer as well to content topic um if a parliament discusses something normally it's only politicians within that on cat commission it's 50 politicians 50 let's say real people but nominated by politicians and i was fighting super hard to to get a researcher in from the field and the other nominee was nominee was somebody um from a big tech company because it's normal that for ever for example the the social democrats would bring in unions the christian democrats would bring in um their stuff so um and then i remembered uh joshua he it was fun basically to work with him and the good thing is and i think that's that's basically bridge to what we should discuss it's super hard for politicians of politics in general to have a vision what to do with that stuff if you re if your
child if you're in charge to regulate it's super hard to find the goal what to do with it and basically this is when the researcher comes into play um to bring positive ideas what to do and especially when you have to regulate something you don't know yeah so it was a new for most politicians it was the unknown we were talking about in this oncat commission and that was really for me also if if i can start with the resume of this on cat commission for me that was one of the the nicest thing to see how the uh how the knowledgeability of the politicians raised in the first few sessions right from having an a faint idea what what we are talking about when we talk about ai it became very concrete for most of them they could they got tangible ideas of what we are talking about and they also very soon understood that that we are not in the age of regulation yet yeah but we are in the age of exploration
and getting to know this technology and so we must not stop it in the very beginning by over regulating it i think that everybody understood very soon i mean some still tried to to to to have red lines and and these things in there but i think that they were getting softer and softer but fabian in his keynote which was enjoying very much said the moment the politicians come in it becomes unsexy so basically since you work with politicians for two years you're as well causing unsexiness so um what was the point or would you agree that it becomes unsexy if politicians come in and if it is uh like favian said why is it i mean you're you're the external one with the real world viewer i'm i'm infected already it really depends on which politicians you talk about i mean we had politicians that are especially um experts for for helping smes for small and medium companies and they were fighting really hard in the discussions to get good uh good requirements in
this in our final report so for example that the small and mediums enterprises that they need certain kinds of help that they need to get access to data and so on and so forth so it really differs i mean the politicians are fighting for their constituency and so no i i don't think it became unsexy so we the funny thing about this digital transformation and ai is that you always have to cooperate with people you need data we need data for example from the public uh institutions we need data from the industry we need the use cases from them we need to get kpis from there so we cannot do it in our scientific ivory tower it's it's no fun that's the funny thing about ai it's not really sexy when you don't have questions and when you don't have users right so i mean often enough users are the enemies my my wife is the designer of course then the the customer is the enemy yeah you can do it better without the customer
intervening but in ai and then i i wouldn't fully agree with fabian it gets more fun when people are using it when there is data and when there is uptake well actually we are now back at the title um we normally build paper tigers and you said basically it's fun when there are customers and so you see i'm becoming a moderator but um what would be your your top three things which we could achieve um basically within that commission from that standpoint from really execution so what will help i mean knowing that we created the paper tiger that's part path of the um reality as well so there is nothing cool nothing clickable it's a pdf so you see basically how um how the the real goal and the execution is drifting apart but what would be the top three things written down in paper the next government has to do to support basically you or research i don't want to make make this paper too small but i was hoping more about the hats
about the people really in this room who are now convinced that something has to happen they can go out to their parties and they can talk and they can explain what's going on and then they can point to the paper and one thing for example we put in this paper i find personally really interesting but it's i mean one of hundreds of of requirements that we put in there so that we need apis for the big platform so that researchers for example and journalists can study topics like uh hate speech or like like political um um uh um what do you call it influencing of of of people so that the journalists and that you can now say we need to have these platforms with these apis we need to force facebook and twitter and so on to give apis to journalists and to and to researchers for studying the phenomena everybody what he wanted to talk about hate speech and these things but we don't know about about social bots and we don't know
whether they exist yet and this was one point where we were very concrete where we said what needs to happen next and this will happen and we also talked about of course that we have to find good conditions for research and good conditions for research research doesn't only mean money and and professorships and so but also need me means kindergartens and means i mean yeah second dual career chances and so on i think this has also been stressed again and and and many other things but maybe as well to the people in in the chat and in the internet so uh shoot us some questions we we don't need to to uh talk to each other all the time but coming up back to the apis um so i mean there are guys out there running an ai company so what's your feeling on that because that was highly debated you know and basically to say somebody just open up your data give us um give us access and if you remember the discussion uh
there were parties uh who said perfect that's something the bka should have so the big r german basically an institution which um i personally would not say these are the ones who need access to the apis so it's it's a fine line you know to split up someone's business and say we need access for you as a researcher it might be nice but for a regulator that's actually a hard shot and i'm not trying to say facebook really not but actually that's a big thing and so if somebody is out there thinking about that alyosha gets access to the apis to see if there are filter bubbles so just share your feelings with us because this was actually a mess when we discussed that's absolutely true i mean some things that are public that can be crawled in principle i mean that we can information that we can find but it gets much easier when you do it through an api i mean this is undisputable right this is no problem but otherwise i mean
we discussed this and this is really an issue some people say yeah the state could could be could help and could do to fund this research but if you look over the border to poland or hungary or the former u.s president or china you wouldn't want to give this in the hand of the state you would rather say there must be a third party maybe an ngo or somebody that is controlled by the state but not run by the state needs to get access to this data and i think um and this was another thing that that i learned in this democratic process that we had here we didn't really pinpoint who should have the access in the end we said we need access and we said we need to find and debate further good control structures how this can happen right we gave examples of how it could happen but we but we uh through this as you said through this um very broad political spectrum that was covered there we we all most
of the time we said this needs to be further debated and we need to find the suitable things that i wouldn't i wouldn't know how how i would prefer it but i want to have the access to study these things but it's as another thing which is interesting here you appointed to europe and we all want to have ai made in europe and as well in fabian's keynote he said europe basically needs to live up to it and i'm i'm fully convinced that this is true but now you brought it up because when we were debating internally and not in our commission but about the corona app for example about central or decentral storage so if you do centralized storage and european syncing this means basically your data goes directly to mr auburn which is probably not what you would want for the 80 million of the german population so it became decentral now everybody for a reason as well is criticizing it's not integrating nicely so even if we always push europe it's a
fine line because i would not say that a whole of europe we agree and so that's basically a problem for a digital single market as well so and that's i guess hopefully something you learned as an external guy that calling in europe is theoretically easy but practically has some downsides as well yeah and at giving technical solutions to societal questions that are open right this is another thing i mean like we knew it from politics when it gets hard then you often get get to the to the legal question right about i mean homosexual marriage then you go to khalsa and then so that the highest court has to decide whether it's possible or not but in principle it's a political decision right it's not a legal decision right but we but it's easier to give it to turn it into a legal decision and i've also sometimes had the impression that we turn some societal questions um do we want to do we want to use face recon recognition in public places that's a
societal question and then it turns out into a technical questions about the current state of the art how many false positives how many false negatives do we have on the current state of the art and technology which doesn't matter to me to for answering the fundamental question whether we want to have this kind of technology you can you could do it the other around you could say we want it as soon as we have the following error rate then we want to use it right and then science has to make it make it work right this this would be for me would be the right way to do it not the other way around oh what is the error rate can we accept it so slightly let's let's see yeah what can what can research offer to me i think this is an old-fashioned idea of transfer it doesn't work here i i need to start laughing because this brings us back to our one other massive discussion without uh almost any result or know
where to dispute is technology neutral you remember you remember that one because now you basically you argued with your face facial recognition that it is neutral is it really neutral or is it does technology have an agenda because the ones who are not in the room this was debated i guess for an hour and then we stopped the debate without any result and basically this topic at the end came up and then there was a voting and now it's some fiddly fiddly text in the in the report document because we could not find the solution if technology is neutral there was really a split in the house three parties more said yes three said no so what is your final conclusion is is ai neutral or does it always have an agenda i mean of course there is a philosophic fundamental uh position that says everything that is made by man is not neutral but but but human values are in the every hammer i produce every car i produce is loaded with values right
it and then it's no more neutral but i mean it has my my intention has gone into it but if we take this away and if we're a little bit more liberal in the notion of neutrality when you it's complex when you think of the the development of an ai system somebody at some point uh maybe for the bachelor thesis has a new machine learning algorithm somebody uses it and turns it into a image recognition algorithm maybe years later puts it on open source somebody maybe years later trains it with images to make it recognize i don't know animals and somebody later turns it into an app that can be used to to recognize animals it's a long development process and the question is when does it lose neutrality yeah in this long chain of things i mean and the person that that for example turns it into an image recognition doesn't know it will be used for animals it could also be used for cars or for for uh complex situations so you don't
know and then often people say ah yeah the computer scientists they need to swear some something or they need to get a algorithmic tool and so you have all kinds of measures that that could help but they don't really help and at some point in this long long chain it loses its neutrality it's not it you cannot say it was then or then but it but it's complex and that's the point and we wanted to make clear that the basic technology most and i i'm convinced that the basic technology is neutral right the other day somebody told me how well already our way of doing mathematics and statistics is highly uh influenced and and other like japanese people they have a three valued logic maybe and so so i mean if you question the basic assumptions that we're using like base statistics or so then nothing is neutral but i think that goes a little bit too far i think we should if we look from the far end uh i would say here well
in the beginning it's neutral but but as soon as you feed it with data i would say it's no more neutral of course when you choose data and feed an algorithm then then the values that are hidden in this data are in the model yeah sure absolutely being an investor looking for returns i'm getting a bit nervous here i don't know about the audience but um so so the topic is a bit aiming at maybe we germans we're slowing things down because we are paper tigers um at the end of the day people who are using ai they want to make something out of it they are creators right um why is it actually happening that we have this paper tiger image what are you guys doing in this commission are you are you trying to prevent those entrepreneurs from doing doing what they want to do or what's happening there what is your motivation i haven't haven't understood that yet yeah absolutely maybe i'll take that i mean you see that's that's was a
good example of actually what was happening there so you're coming from the field side so i was in tech before and i would say the political side comes from that that questions we just had because if you want to regulate something if it is neutral you would end up with another regulation then it has an agenda so now i'll stop that because this will go forever and ever but to make you understand what the problem is you cannot just come up with say um look we need we need a form for startups whatever had an innovation fund then somebody would question what should it do should it help basically science should it help common wealth should it change our social system so the real problem is what i learned in three years in politics everything has to be attached to some political sense or goal and this is absolutely not matching to what's going on in the field and to give you some concrete basically winning so while discussing the problems um to basically get
funding from startups with all the parties over and over and over you could see that the the current political direction was changing so the government is now talking about an innovation for which should start next year i think it's 400 million from the government 600 millions free money basically flowing in from you from investors from companies and this is a result that they realized and they really had to talk to startups with that problem and now this for is coming up without being attached to a bigger political meeting it's just trying to help startups um to survive basically the the second level second investment wave or something so but this took i think i don't know three months to decouple it from political agendas and that maybe explains where where the problem is and why from from a field perspective this looks like we are all doing some esoterical stuff and as you said these political uh topics they are there already right then somebody wants to talk about dsk4o and other people want to
talk about future of work and this and that i mean right why do we want to talk about the future of work well why is the problem of the commission on ai to to discuss what happens if if uh just in case i i would would kill some jobs right so you could also to take it you said that's not that's not our point here we want to help the technology uh get through and get germany uh become first in the technology development and then when it comes and when there is a job loss then we could start discussing what can we do about it if it doesn't come we don't need to discuss it at length right and and nobody knows as fabian says in the beginning whether it will come so i think there is also there is also some some some mismatch but i think for some people we could help a little bit with their fears also about killer robots and so i mean this is another thing that some people
also think that when you invest in ai um it's like an atomic power or like genetically modified stuff yeah once you have it you will never get rid of it again so better not try it i think that's still in many many people's heads and this is also something we have to work on and there mario you said it already and you said it in another at another place we felt a little bit short in the outside communication but also because it wasn't the um the the mandate of this of this study commission right but this is something that needs to to change the attitude of people in germany but maybe bring our investment moderator in what would be your expectation in terms of that so again i'm i'm really i'm in favor of what the government is doing there which is absolutely not my job as an opposition that they learned that they should come up with an easy and flexible investment for but still you could see in the next level of i
would say being open to it institutional investors are still not planned so meaning that our pension for money or something could go in that instead of investing into concrete or whatever um they could invest in the next generation why isn't it because still the political um i would say scare to say we invest some pension money into something which might not be super safe is not there so it's not yet there and we could not solve that through the discussions in the commission with ai and some examples because basically their the political hindering was still too high so we would need some some in business you would say storytelling from the field to convince the ones who are scared and if we are honest that's still not there because fabian said before ai was dreaming and now we are in execution so but within execution you need political visions to put money there to basically get it kick-started and i guess we are a paper tiger because sometimes we we at least we're germans are
not that visionary compares to the americans they sell us a pen and had they have a slideshow that you think oh my god i need to buy it for 400k and we say it's a pen it writes you know we are too rational and this comes back to um what favian said if politicians come in it's unsexy um but i'm trying i'm really trying to change and i hope that it's changing and fabian as well luckily said he thinks it might change i mean this is maybe all things some one thing we put in this report is that we said also in the educational sector we have to be very fast and we have to bring computer science to the high schools and everywhere i mean i did my high school diploma in 1994 and i had computer science lessons already back then and still today people can do their high school diploma without having ever touched a computer would you agree to bring ai then in like recognizing the eye tracking and see that
you have reading issues because that's that would be the blocker topic so i'm now the politician i always tell you why it's not working so convince me uh if you do it right i mean the question so mario's question is when we put an eye camera in the classroom that for example can measure whether the pupil whether they are attending or whether they fall asleep right in china this happens right do we want to have to have the same in germany and i would say if it's only an immediate feedback to a teacher or say like a university teacher in a classroom with 800 students sitting there to have a meter that tells you how many people are now sleeping or or watching they're watching movies on their iphones that would help but if this information leaves the room and get goes to their boss or maybe to their lawyers or their parents and so they can yeah so they can they can misuse this information uh against this teacher i mean then i wouldn't
wouldn't agree but it's it's a question about how it is done and what happens to this data yeah is it just an immediate feedback that will be deleted or will this data be saved i mean the same we have with uh with uh saving telephone calls and so on and an old discussion right that we have for a long time you have to make it right to make it happen but but but this is really um you see but here so let's come back to the to the paper tiger thing and how we fixed it because for that example it's interesting so the parents don't want it which then we would be back to the argumentation why the customers don't want it and um in that end we are not in but um to again give a positive example which i at least think was was helping as well the guys with companies in the field the researchers um before germany was defining ai excellence clusters by a federal state so we had i think
16 because everybody got one so that was the political solution now we are down to five um so you basically as a researcher you hear there will be more money to attract talents um even if you lucky lucky you the the dfki is basically not bounded to the bessa stellung boat so the ones who are not german there was a law that as an ai researcher at for example dfk you could not earn more than somebody would earn with a state contract in whatever state office so and this was highly blocking talents to come in so do you think five excellent spots in germany still realistic or is it still a political wish compared to the size of germany and if you think how many excellent sports whatever you can you can count off in the us or in china i mean if we don't raise the talents it will be hard to feel the positions the same we see with these hundreds i mean professorships that are floating around i mean one of the
only kpis from the german ai strategy right and everybody's asking what's happening i mean today it was in the newspaper again i mean so that we have trouble there but i think if we find if we build the talents i mean why not why not go for for uh for for this number i'm not for an absolutely centralized one hotspot i think we can do it we can have several of them that's also what we do at dfki and so they can build several um topics for example we have at the fki office in bremen close to the sea and they are working on marine technologies right they are closer to the sea they wouldn't do that in munich yeah but in munich you're closer maybe to car industry and you can you can have a hot spot there or like in the stuttgart cyber valley region people people do that i think that is okay to if you link it to something sensible uh then you can have a little bit of decentralized structure
then i would basically stick to the decentralized thing because um i think we are actually missing out on that too much so fabian said we will have our own facebook i'm not sure fabian if you listen to that and i'm not sure if we really need it and if we would have it i would hope it's decentral so maybe talk about um because always we talk about china central stuff blah blah blah and there was always the question uh can we live up to that and are they too powerful so i would just want to say when it came to car or whatever um electrical engineering and that stuff we were never bigger or richer than china but we were just better so i'm i'm always question if it really is the right way to try it to get on facebook and stuff or if we should focus on really just becoming an active network so maybe you guys on that i think the big big to be things they have been eaten right i
mean that's the that you have you have it in the us they were the first ones and then in china of course you close the market and then you can replicate it once but we don't want to close the market in europe and have our own facebook and and uh and what have you itunes uh so i think we have the b2c we have the strength in the small and medium enterprises right so we have a tax office in dortmund and we have a uh car manufacturer at uh in stuttgart and they want to work together they want to work out some nice um new kind of business that they have to i mean in this b2c area where we also we are not talking about anyone let me just run because eduard said it's not about the money it's about uh speed and progress so absolutely i'm in so but then then it's about willingness of a society to change so basically we are back and i don't want to open up the political
discussions again so um is our society then hungry enough because in the end i wouldn't say china is that the chinese people are not that hungry they are at least forced to be hungry but are we not hungry enough then i think we have a negative attitude towards technology i mean since 30 40 years so every new wave of technology came with a lot of protest from people because they thought somebody wants to trick me into something for making profit i think that's the general attitude we we see a lot in germany and i think it hasn't changed now we don't see that the the positive potential of of this technology also when i read the newspapers i mean we get these paper clippings and i've really carefully looked through our two years of working on this commission does the the the the the the do the messages change but i think they haven't changed yet the other day i talked to a young student about uh ai platforms that should help students do their
the do their university get the right courses get the get some feedback and so on and they said oh this is total control now they want to to control us uh that's that's the feeling also of young people but maybe then to answer eduard to give you the positive thing we were debating a lot to get in the um the so-called regulatory sandboxes to basically not hold up the ones who wouldn't be faster it was basically super tight to even get them in as an official recommendation because there was scare already and i found a nice article they counted the words chance and risk and it was two times more risk than chance in it and remember how we were fighting so basically um i hope we at least we could bring up these wormholes for the one who wants to go faster so i'm i'm not thinking that we can at least change the ones that are here so the next decade might be a bit more eager to do some changes but it
really it's about bringing in wormholes for unbureaucratic execution and so they are in and basically whoever will be the next government must then argue why they are not fully executing because it was every fraction in that house agreed to have that as a recommendation and therefore that's what i always explain because we are running out of time so in in the in the industry you would write a moonshot paper that's just not possible we were writing something which is pushing from the bottom so we were basically pushing up the the minimum bare limit of innovation which by itself does not sound fascinating but in the end it is because there is everything what you will find in there there is no logical way to not execute that because basically it was common sense at the whole parliament so maybe if we read it the other way around then it's more uh refreshing i would say so and we are running we are also maybe a little bit of time don't worry we have uh yeah
we are maybe the wrong wrong people to talk to because you are from a liberal party i mean you said let let people try it out on the market yeah and the good solutions that will survive and i'm from research i say okay let's test it out of course i mean do it in an in a human-centric way i mean have the users on board and have the trade unions on board and everybody that is that is um uh um inflicted by this technology right and let's make it good and let's make it right but the but it but i know i mean everything will be will be worse when we don't do anything that's the point i mean there is no uh there's no guarantee that that everything remains the same when we don't then let's put it back to you and maybe our moderator as well as investors so you're happy to join and i was confused by a big zeros in red looking up my screen therefore i thought we were running
out of time um did then you as researchers or even as investors in company um did you miss out to explain nicely what you're doing because i mean there was the point talking about explainable ai it came up all the time but you really look if you really look underneath the hood and all the research project it's not yet that much so if i would put myself into a classical political standpoint so did you miss out to explain what you're doing i mean there were ai windows it was there all the time so why does nobody understand it then the white white the my question would be why do people need to understand it right for me it's the other way around people should to tell me what they need and how i can find out uh if it's good when i build it and then i build it then they can test it i don't know that they don't need no not everybody needs to you don't need to understand how the brakes of
your car work exactly it's a lot of high tech in there but you don't need to know it you just need to need to be able to operate them and for for for for a doctor for a clinician it's completely different when a system gives a recommendation for some some medical therapy then they need to get some sort of explanation but they need to get the right one right and they need to tell us and that's what we're doing in projects then they tell us we would need this and that information to to understand whether the machine understood the case and got a good recommendation right but this is very particular for certain applications and you cannot do once and you cannot discuss that once and for all and that is something that we also found i think everybody understood in this that we need to differentiate yeah are we talking about mobility are we talking about medicine are we talking about entertainment that we need to that we need to act differently there but
i mean but you're i mean as a side remark now i'm talking really as as a scientist uh all this about explainable ais is really is really a sad thing for me because i started out in the old days of good old-fashioned symbolic ai you understood perfectly well what the machines were doing but they were very simplistic a robot could lift their arm and could pick a box but you would understand why they would do it right and then came all the machine learning and the system started to learn themselves the performance went up through the ceiling we know that but we don't understand how the machines do that i mean this is also very disappointing how can a machine translate text so well when we have no idea how they do it internally right i mean this is really disappointing for us as a scientist because you want to understand how things work so i would be the first one who would she allowed if we understood better what the machines are doing inside
investor mario it's pretty easy to be honest um our companies look at things that are working and and as arya said so it's really testing and trying and you're not really asking the question what is the politics what are the politicians thinking about this what is regulation thinking about this we were actually a bit scared when we heard that there would be something that users could claim like later that you take their data out of the training of the model and would have to kind of get the model to a state when the user wasn't there because this is highly complicated but i have four i'm sitting on boards of four b2b companies so that topic is a bit different um so it's not like with end consumers it's it's in a b2b context so still gdpr is very important and i think by the way that's a big strength of europe that we can really ensure user privacy and we can enforce that also on other companies that are coming into europe so this
is an edge of of ai companies starting here plus and i think that's that's the second one um it's it's really about multilingual approaches i think ayosha is more more firm in that but two of my companies they are in nlp um and and they are the only ones who can do it because for americans for example they they speak english and that's how they train their models um so but at the end of the day they don't want anybody influencing them they want to be creative and build something that people buy and and that makes people absolutely so don't get me wrong don't start to think what would the politicians want no because then we are doomed in the end so that's that's absolutely not um how it should work on that so don't get me wrong but maybe to end this a little bit because we're now we're running out of time uh since christmas is coming up um i would love to take the opportunity to for both of you to ask
each other what you would wish from the other side to do so the scientist asking the politician and the politician asking the scientists to make really the application of ai from a german perspective much much better and faster who wants to start i could start i mean one thing i would really hope is that that the state itself would start using ai for various purposes and telling people what experience they make so that they are really a forerunner so that we would say oh yes we have a cool state and now the companies they cannot stay behind they have to be as cool yeah so when i when you apply when you do your text regulation and so on and so forth i mean then you could get a nice interface where you can where you can put everything in and you see the effects and so on i mean this would be really nice and for many other things that are paper-based or fax-based still that would be the one thing and at this
at the other hand i mean of course you are no opposition not not in the the current government but it would be good for me also if we could could a little bit monitor the long-term effects of our paper tiger right i mean we will not see the effects tomorrow or the day after but it would be good if you would call call on us i mean on our expertise i mean regularly so that we can monitor and can see if we if it's still their recommendations are still good you have some of them have been met and they need to be modified i mean this would be something would be good i think to have a little bit more long-term activity than what we could reach within two years thank you mario so um basically it's really research communication um i think it was missing a bit or not only talking to the politicians but talking to the people so you're you're the bridge in between and so i mean there are minor people
who don't believe in research anymore but i think this triangular thing needs to work better because you come without an agenda and if you see um what happened within the the covet crisis so radar if you share the opinion for example of mr truster or whatever so what they brought up how many people listen to the podcast and call us and say yeah but he said and we need to become better in our argumentation so i think this this circle was a bit sleeping and therefore scientists should communicate a bit more to ordinary people and not only on conventions and stuff because this really helps the politicians to hook into something which for voters comes without an agenda because otherwise you're always the politicians you want to do it for this in that region so this would be my christmas present and a bottle of wine that's a lot to do great thanks so much thanks a lot both of you mario yosha that was really really interesting um and i guess if we have
more questions later on they can just shoot you an email easy absolutely cool thank you so much please stick around and everybody we're not taking a break until 2 pm and then we continue with ai strategy so hope to see you then again thanks bye-bye thank you