good evening ladies and gentlemen welcome i also would like to thank the two other rooms which are connected it's interesting that there is such a crowd here he will speak english so you might need the headsets i will make a short introduction then we'll get the floor to our guests and we will devote a lot of time to q a so that we can have a discussion together uh he's a sociologist a writer a journalist from white russia his article and his books have studied the impact of new technologies but also indicate some line of reasoning why while connecting the advent of new technologies and with the geopolitical scenarios how new technologies are used why it's so difficult to analyze them without their context so his books have always been a very important reference point for technology enthusiasts will speak about artificial intelligence ai in china and the us and the comparison between the two powers the signs are already there because there are immediate consequences of that you might have followed the discussion concerning uh custom duties and the desire of trump to punish for the budget and the trade surpluses at present the duties uh are about 25 and they are scheduled to be implemented by mead june and the idea is to have an impact on in chinese production china has invested a lot on big data and ai in order to challenge the position of the us all over the world i will speak about china for i have lived there for about 10 years and have been you know witnessing the uh changes that have happened in the past the uk factory of the world an economic model uh highlighting the low cost production in southeast asia that had allowed the growth of china after the reforms of deng xiao ping but change was already there in china and in the year 2008 the olympic games took place and after that china had realized that there was a world crisis affecting the west in particular too um and they witness a drop in the orders so the attempt to modify the economy from an economy drawn by exports china tries to be drawn by domestic market and that is because of internal mechanisms because of the process of urbanization and exponential growth of the middle class and the attempt to turn quantity into quality that meant leveraging on ai technology big data and innovation there are some moments in time which really were key in terms of legislation first of all the project by president xi ping which changed a lot the international standing of china till that time everybody was indeed very suspicious but then after that they took the lead in globalization also uh orange of the isolation policies by trump and last year the davos forum was inaugurated by the chinese president where he defended fiercely globalization we have a project called china 2025 hinged on the idea to make china an economy drawn by ai big data and robotics and trump is there to strike on that project through sanctions that to tell you that this challenge will go on for a long time but the year 2030 china is determined to close the gap to the u.s another project is that called internet plus which concerns ai i'm not going into details now because i would like to give the floor to our guest without further ado but first i'd like to tell you that besides the two projects in china there is a legislation asking all foreign companies collecting data to live them on a chinese server and this is mandatory so there is a clear attempt to control on part of the government on ai and big data on the one hand this is done to favor the export or high-tech products don't forget that uh us the us buy robots from china mainly and also because domestically speaking the mantra of the communist party in china is to maintain stability and then nothing should change in the society in daily life life the search for ai is a fact china is uh achieving our dystopias i might remember the black mirror uh program where people were assessed based on their social score this is a system which is already there in china so your position in the society depends on a number of elements first of all your criminal and administrative behavior some tried to buy a a ticket and they were refused because they had not paid for a violation they had committed 10 years before that might also be connected with the use of e-commerce instruments and so on there are projects launched by companies like alibaba which had created a scoring system called sisam point used to trace our online behavior on purchases in order to guarantee loans and to be on the safe side to know that people could settle their debts they also went beyond they wondered you know that person is he or she buying chinese or foreign products uh are there butchers or not the chinese central bank stopped these projects saying that that was too much but still the idea is on the air but not only that in china many shops can be paid through face recognition last week in anzo school near changai a pilot was started with a tv cameras analyzing the behavioral students in the classrooms you know there are a number of examples which are already there the most advanced are happening in the most sensitive regions for the governor in the xinjiang northwest majority of muslims but yugori who used to be a majority then they were chinese eyes as it were and the police chinese is already using predictive models through cameras there will be one tv camera every three chinese in the future with intelligent smart cameras they sold them to the us and an attempt was made to sell them to italy but the project was rejected through face recognition where that are matched with other data banks were some predictive models like minority report so they are implementing our dystopias model has his own position about the ownership of data he uses the adjective common that means they should be owned by citizens china is proposing their own model which is not the american one as we saw with the case of cambridge analytic which was a private company there will be a government control of data collected by chinese and foreign companies which will be forced or rather mandated to live their data in the chinese service and no company will give up possibility to work and market their products in china so since the year 2030 there is a very precise plan to uh become better than the us in ai and big data but this challenge is uh carried out in terms of uh trade influence but which might give rise to very strong competition and that will be the um play game where the development and the future of artificial intelligence will be played so the floor to our guests and then will entertain the questions thank you well thank you very much simone for this very good and very in-depth introduction also to the chinese part of the story which i'm sure you know better than i do so um let me try to give you an overview of what i think the main themes of this unfolding debate about artificial intelligence and geopolitics would be um and i will try to leave some time for questions uh for you and maybe from simone at the end but i really have a lot to tell you so if it gets too fast you have to wave at me because i tend to speak very fast so before we get to the heart of the matter let me just be clear that when i speak about artificial intelligence i do not refer to the skynet you know i do not refer to the terminator like artificial intelligence that awakens and takes over the world i'm talking about something much more banal and i think that's also something that most people who analyze these issues from the economic side and actually from the commercial side have in mind when they speak about artificial intelligence we're talking about still relatively basic and primitive ability to do relatively trivial tasks recognize certain objects one from another in photos be able to recognize voices belonging to different people being able to analyze photos and write a summary of what the system sees in the photos so they might all look rather trivial to us because it's not the kind of full-blown artificial intelligence that mimics the human mind that we were promised in science fiction novels however i would argue that taken together this functions however trivial they seem have an enormous uh influence an enormous impact on economic and political activity which is why everybody is so busy uh discussing them and they're not offered uh by uh you know some kind of startups based in the garage even though there are quite a lot of startups working in ai there's probably an area with the most startups uh uh you know square meter in china or in silicon valley but the majority of these services as i talked about them and i think as most you know financial newspapers like financial times wall street journal talk about them they're offered just by a handful of firms four and five of them american and three or four of them chinese right so we're really talking about a relatively small number of players that have amassed this ability to offer this relatively trivial services recognizing objects recognizing faces recognizing voices and so on right so we're not talking about every single startup in beijing or in san francisco we are talking about the usual suspects facebook google amazon microsoft ibm which is more or less the american part and also on the chinese part of course we are talking about companies that are somewhat less well known in europe but they are almost as big as the american ones alibaba baidu tencent jd.com and a few others um before i walk you through some of the dynamics of competition between these firms uh within the american market and within the chinese market and what's driving them i think it's very important to understand the exact point in the kind of arc of digital capitalism that we currently inhabit right so i would argue that we are in a very particular point in the development of digital capitalism and that's a point that is quite different from what came before uh and this is where it's very important to grasp one particular transition so up until roughly 2014 i would argue 2015 the dominant moral of many of the firms in the united states to some extent the firms in china was online advertising and in the case of amazon it was the sales of products which most of us bought from these firms it was primarily oriented to consumers buying stuff so consumers buying stuff had to see some ads telling them that there are good products to buy and they had to actually go and buy those products somewhere which many of us did on amazon at some point in that in the development of those services some of these firms especially amazon or also google they have recognized that there are many other domains and areas of doing business that are even more profitable than this consumer-facing business that they have been developing for decades so instead of just selling goods and service well mostly goods actually to consumers advertising goods to consumers you can actually start selling services to businesses and to governments and if you start selling services to businesses and governments instead of selling and advertising gadgets and all sorts of other products to consumers you might potentially even make more money with it amazon was really the first one who grasped that development early on because in around 2007 2008 amazon realized that it has a lot of spare capacity it built a lot of infrastructures it built a lot of server farms it built a lot of basically unused capacity which it needs in order to be able to sell its products and it realized that it can actually rent out some of that capacity to other businesses who need to store the data who need to store some of their sensitive information who need to run certain services on top of that infrastructure and so forth and they've discovered that up to a point where even cia and much of the us defense establishment now hosts its data on amazon servers having discovered that and that by the way we're talking about the shift to cloud computing right it might seem that i'm talking about something relatively banal or something really reverb score no i'm talking about very profound transformation in how many businesses structure their in-house i.t basically more and more of them instead of just buying servers storing their data inside the firm taking care of all of it they now just outsource it almost completely to firms like amazon and more and more firms have jumped into the mix because they have understood that amazon while it was not making a lot of profits not a lot of revenue but a lot of profits on selling goods was making very high margin very high profit margins on selling services so all of a sudden google a company that primarily was selling advertising also started selling cloud computing services microsoft started doing it even earlier you know a company that was primarily building software that powered our computers suddenly became a big player in cloud computing so you can see that despite um the current framing of this digital capitalism is being driven by monopolies which to some extent is true because google has monopoly in search amazon has monopoly in uh selling products these firms nonetheless still compete in some areas which are very important and those are also areas that they happen to be making most profits because even google makes more profits and has higher profit margins on selling services than a dozen selling advertising and it's very important to understand this transition from selling their primary goods advertising or products in the case of amazon to this idea where they're selling services because once they start selling services as firms you can see that the sky is the limit in terms of what kinds of services they can be selling so if in the beginning they were primarily selling cloud computing services now amazon offers you more than 100 different services through its amazon web services package and those services include everything you can basically upload an mp3 file and amazon will automatically transcribe it for you and you can upload so you essentially get a text out of an audio file you can upload a photo or several photos and amazon will recognize what objects are in that photo and will tell you that this photo goes in one folder that photo goes into another folder basically all these simple three-wheel tasks that i was describing to it very beginning now they are offered as a service right whereby you pay for the usage of that service so if you upload um you know a text of 500 words and you want amazon to generate an audio mp3 file in a human voice done of course artificial intelligence and then you want to download that file and use it somewhere else amazon will charge you peer word so the more words you have the more it will charge you so for many businesses this has created a very useful model whereby they can offload even more of the functions into the hands of those firms so we are not just talking about them leaving their it infrastructure in the hands of the providers of cloud computing we are now talking about firms leaving functions that previously had to be done by secretaries by clerical workers maybe in some cases by creative workers all of those tasks are increasingly being farmed out to firms like amazon google offers the same set of services and so does microsoft chinese firms of course are clearly driven by similar set of incentives even though in the case of 10 cent and in the case to some extent of alibaba they managed to tap into other sources of revenue that were not advertising very early on because in the case of 10 cents there were many online games there are many other sources of funding serious chinese firms could actually become so profitable so they were not as tied to advertising as silicon valley firms but in china nonetheless the dynamic is very similar in part because increasingly american firms in chinese firms compete for the same global market so the fastest growing company in terms of offering cloud computing and web services this year and last year is alibaba right so we're talking about a chinese firm that has come out of nowhere and is now competing with essentially google and amazon and microsoft offering precisely these kinds of services that i have been describing to you now and all of them are essentially going after not just the their own respective domestic markets but they're also going after international markets offering their services to clients in russia latin america the rest of asia africa and so forth right and they're offering it not necessarily to consumers but they're offering it also at enterprise level and at the government level the reason why i emphasize this transition so much is because it's very important to understand that the transition away from online advertising as dominant mode to essentially collect data right and to essentially offer many of the services and have immense repair questions or also how the internet quote-unquote looks like today because you have to understand that in the process of sticking to this first stage of their development where all they were doing was working with consumers and selling advertising and selling products these firms have of course accumulated huge tons of data not only have they built immense data centers which amazon then managed to rent out as infrastructure those have managed to amass quite a lot of personal data and enterprise data which later proved extremely useful in building some of those artificial intelligence services and that of course is what paid for many of the free services that we have been enjoying so the reason why we pay nothing still to use google or to use facebook it's precisely because as all of you by now should understand there is an implicit extraction of data from all of us which then is monetized in different ways we give you a quote from andrew eng who was the head of ai at google and was then also the head of ai at baidu so he had kind of been in both worlds and speaking at stanford last year he said something that i think is very key to understanding the dynamics of the digital economy he said that in big tech firms we often launch products not for the revenue but for the data that they allow us to extract and then we find a way to monetize their data differently which i think is a very honest um explanation of how the digital economy functions meaning that all of us are getting in a sense a massive subsidy to use google and facebook to some extent amazon for free and then with the data that is accumulated these firms find ways to actually monetize it at the beginning it was advertising now it became services including services of artificial intelligence let me give you one example so and that's an example that i think is used very often that people don't understand the full chain of implications here so every time that you use google when you use it in a way that google finds suspicious right it will of course ask you to prove that you're not a robot and that has happened probably with all of you and whenever google suspects that you are a robot it asks you to do some tasks right and most often it's a task which begins by displaying you nine photos taken mostly in urban settings where you're being asked to find you know all the boats that you see in the picture all the cars that you see in the picture or to identify all the you know front shops or something of that kind right and you should keep clicking until you finish and once you've done that you move on you get access to your content as of course most of you probably i hope so have been suspecting that's a very clever way for google to teach its own ai systems how to do object recognition because essentially they have deployed plenty of cars to go around the world and take photos of the urban environments but the cars themselves of course don't know what it is that they see and deploying the ability of hundreds of millions of people if not billions of people on daily basis to analyze those photos and tell google that this is a car this is a boat and this is a street light it comes in very handy because it ultimately helps google to develop image recognition and image identification capacities which they can later insert into other services which they will sell to commercial and government clients so if you've read the newspapers carefully last months you must have seen that there was a big protest inside google led by its own engineers whereby they were actually rebelling and telling that google should not be working with the pentagon because google entered into a deal with some of its technology was used by the pentagon in order to basically recognize objects when it was operating with drones because drones also need to be able to recognize objects not just self-driving cars so indirectly every time we teach google that we are not a robot right we are enhancing the capacity of institutions like the pentagon to identify houses in yemen that they're going to strike through the drones right which of course is not something that occurs to all of us but it's something that's immediately monetizable by google and when you can deploy millions of people to work through for free to be able to identify these objects in those photos and then you have the ability to sell that service to an institution like the pentagon you have a pretty good business model in front of you now there are many implications that flow from the shift from advertising to services which i'm not going to explore today because we want to focus on geopolitics and ai but i think it's very important to understand that one of the consequences and questions that we should be asking is who's going to pay for the provision of all these online services right once these firms no longer need us as providers of data they're just something to put in your mind because clearly they would soon arrive a moment where google amazon facebook and all the rest of them can make much more money by essentially dealing with governments the pentagon banks in the industry than by continuing to serve as ads right and when that moment arrives they have very few reasons to continue subsidizing our use of those systems so that's just something to put at the back of your mind so the reason why china but also a couple of other countries including by the way russia putin made a very big uh remark last year when he said that whoever controls artificial intelligence is going to control the world you know it's it's not it's not just china right so you have more and more countries suddenly waking up to the possibility that the economy nationally and globally gets restructured in a way that more and more power concentrated in the hands of technology firms these technology firms could be american they can be chinese they can be idioms they can be japanese it does not matter but there is a cluster of technology firms that suddenly intermediate many other sectors just look at what google intermediates right they offer services that analyze healthcare data in order to predict in all the occurrence of kidney disease they analyze energy data inside their own data centers to cut their own energy costs by 40 percent they analyze traffic data in order to be able to predict uh when to turn on you know traffic lights and when not to they analyze all sorts of other urban data because they're also building their own smart city right so there is tons of spheres and domains of life that are suddenly intermediated by these cluster firms and these firms as i have explained you know their business model in many cases is very simple they just come extract as much data from whoever they happen to be working with then they repackage that data and offer their services somewhere else now if you think about it from the perspective of a country that is a developing country that is trying to grow its own industry perhaps information industry if not heavy industry and essentially come out of a world that is rather rural still predominantly and based on a very different economy but still very much underpinned by agriculture of course you would be very preoccupied with who controls this key gateway to economic and political activity in your country right and if you look at the history of american interventions in global trade policy over the last i would argue 40 plus years what you would discover is that china is not the first one to wake up to this danger in the 1970s if you go and read many of those debates quite closely you have a lot of developing countries you know it's a so-called group of 77 the countries of the non-aligned bloc who become extremely concerned about the fact that big american computer firms like ibm come to their countries essentially destroy all the local competition establish themselves as the primary providers of the most important services do not want to store any data locally because they argue that that increases compliance costs and then essentially exit whenever it's used them or whenever the profit environment changes so of course a lot of countries have been trying to make it much easier for themselves to a build their own computer and information industry b impose restrictions on what they can do with data and particularly data of their own citizens which is very often simply extracted from those countries and taken elsewhere and they have been defeated in those efforts they've been defeated in those efforts very often by very powerful american companies like ibm and they've been defeated in those efforts through trade treaties right which continuously for the last 40 years pursued the objective of removing any barriers to trade in services and by extension removing any barriers to trade in data so one of the big slogans of you know free trade campaigns in the united states has traditionally been that we want to support free flow of data because free flow of data essentially allows our firms to set up business anywhere not be faced with any local constraints on its use and essentially repatriate at home and work with it according to the u.s privacy legislation which is not often very strict you compare it to the european one uh and that has been a very consistent stance that virtually most countries including by the way the european ones had to more or less comply with until china has managed to get enough how should i say strategic insight into understanding a the role of data and its connection to artificial intelligence and b the need to control that crucial layer of intermediary activities that is now dominated by the tech firms and as soon as china understood it they have been taking very strategic clever steps to a grow their own domestic technology industry b do something about data legislation so that the data of chinese citizens would not be enhancing the ability of google to sell its services to the pentagon and thus make a lot of money on the data that does not belong to the citizens of their own country and c to essentially do everything to engineer with very clever industrial strategy a domestic tech industry that would be able to compete with the american one of course it took quite a lot of power and quite a lot of let's say negative reactions from the west because ultimately one way to do it was to essentially cut off access to big american tech firms in china so they had to embark on a very massive censorship spree where the likes of google and facebook have been banned but not only because the chinese wanted to restrict the political activity happening on those sites but also because they essentially understood it as an instrument of trade and not just an instrument of censorship so for me the great firewall of china is primarily a trade instrument as a way to protect and nurture their domestic industry and only secondly an instrument of censorship and having more or less reading themselves of this big american firm they have managed to cultivate their domestic ones but also do it in a way that is tied to a broader strategy of coordination between sectors between industries and between what each of these firms can be contributing to the overall development plan of the chinese government so if you look at what their firms at task was doing all of their big firms now has a particular theme a particular topic that they sort of curate so alibaba is tasked with uh implementing the strategy on the smart series baidu the search engine is tasked with implementing the strategy on autonomous driving and sell autonomous vehicles and sell driving cars 10 cent is tasked with improving health care services that are based on data and ai all of them have something that is quite specific and they're tasked with it by the government now we can debate what extent this is a healthy moral and what political consequences it has but it's much more coordinated and this much better targeted with the overall strategy of national development than anything that you can see in any other country now of course um i might have given you a somewhat misleading picture in talking so much about data because artificial intelligence and especially deep learning they're not just about data right you also need two other rather important components one of course is people you do need to have quite a lot of smart ai researchers and students and academics who will be contributing to the development of your own domestic effort which is something that is very hard to do uh in a country that might not have had so many graduate programs in artificial intelligence and second of course you also need hardware in a bubble you need chips and because it's primarily by having a robust chip industry and semiconductor industry that you can make a lot of headway in building your own domestic ai capacity so in the case of the united states that domestic ai industry has of course been developed as the result of the cold war and in the 1950s 60s 70s and 80s you have seen companies like intel uh and a little bit later on in the 90s company like nvidia which essentially managed to fill in uh the gaps in the microchip sector so in america the situation is quite good with regards to microchips and it's quite good with regards to data in so far because of american universities is also quite good with regards to human talent now if you look to china their main hurdle is and remains the inability to essentially build their own autonomy with regards to chips right and that explains much of the recent activity and much of the reaction that we have now seen from trump because ultimately the only effective way for the chinese to do something with regards to their chip dependence is to either heavily invest in their domestic chip industry which they have done by creating a national chip fund which has 25 billion dollars to spend on subsidizing various uh acquisition deals with regards to chips and b is to essentially go abroad and purchase as many cheap companies as possible right and then somehow integrate their know-how into the chinese system and the chinese have done it with some other industries if you look at robotics where they just go and purchase entire robots mostly industrial robots and firms they have purchased quite a number of very big important firms primarily in europe which did not really have much um discussion about it and they don't really have much instruments through which to protect itself so the chinese came to germany they were one of the most important robotics firms they came to italy they bought another robotic firm they have bought quite a few robotic firms with chips the situation is not as simple because many of the big chip firms in europe have already been bought by other foreign investors in the uk a very big firm got bought by softbank which is a huge player in the technology sphere which i hope i might return later on in this talk on questions and now you see one of the bigs that one of the big dutch chip makers called mpx is about to be purchased by the americans qualcomm so the reason why trump blocked one of the largest m a you know emerging acquisition deals in history which was a takeover of the american qualcomm by the s then singapore-based broadcom was precisely because americans having understood what the chinese were doing in the microchip space they were afraid that their own crown jewel well one of the crown jewels because those have intel and nvidia qualcomm will essentially be starved out of r d money if broadcom purchases it so trump blocked it right and demotivated quite explicitly it was the reference to china which is trying to assert its power in the space i still think that the future for the chinese efforts here is quite bright in that there are other hardware firms despite the fact that america is doing everything to essentially disable and destroy them you know both zte and huawei i mean the under intense pressure in the united states huawei for example is still nonetheless proceeding with its own plans to build its own chips right and they have introduced actually their own chips in the phones and they made 10 phone there is a huawei mate chip and it's one of the few chips that i made in china with regards to many other metrics china is doing exceedingly well and it's doing exceedingly well precisely because there is just so much money pouring into ai from the central government from this big giant internet and hardware firms that are all very profitable and have a lot of money to spend and there is also coming from regional funds from regional towns and councils and so forth there is also massive pools of money there which explains why china is doing so much better even in the united states and even certainly than europe on most indicators number of patents granted in the field of ai a number of academic papers published on deep learning and ai in top journals in a number of startups funded the number of funds attracted all of that money is flowing to china right and it's flowing to china precisely because the environment despite the seemingly authoritarian outlook of the country and seemingly you know seemingly rigid nature of its legal system and political system the environment seems to be much more flexible and fluid there are not many laws that are enforced about privacy so there is quite a lot of leeway that these firms have in doing with the data of the users whatever it is that they want to do with it which is not the case in europe and is still not the case to some extent in the united these firms st many of the niches and provide many of the services that traditional actors economic actors just do not provide i mean so there is a reason why most payments in china especially in coastal cities are handled by you know tech firms why because the banking industry and the consumer banking industry and the payment industry just wasn't there to do it so if you go to china you wouldn't even be using cash now to pay for most of the products and services that you are buying in part because everything can be paid for with alipay and many other systems that are directly integrated in your phone uh wechat and so forth in many of them you can even now pay with your face because again this is another kind of development that you can only get in china because they have two big firms so the largest startup in facial recognition right now is in china it's a company called sense time uh and there is another one uh of similar size i mean sense time is uh valued around four billion four and a half billion dollars so it's not a trivial startup and the reason why there's so much ahead of the american let alone european competitors is because the chinese government let them access its own database of 700 million citizens in their photos so when you have access to 700 million citizen photos uh of course you can do quite a lot of advances in facial recognition which is what they did and now they go and cross-sell and cross-license uh their services and facial recognition to all the other parts of the chinese economy i would argue that you know if you look at for example their own uber and this company called didi there it's already integrated so this facial recognition technology is already integrated in their uh in their services so what we're seeing is a very different model with much more money and much more coordination at all the levels and just to give you one example of how seriously the chinese take it uh last week they introduced a textbook for high schools about ai which you know i can bath in the next few years it will become obligatory uh an obligatory text in all the high schools where children from the early age will be taught the basics of ai and what it means in part because you know the government has recognized what important factor it will be for the development of the economy as a whole um so speaking about the likely sources of conflict right like the hurdles um of course what trump is doing now you know with protectionism and uh kind of a very retaliatory response to some of the chinese companies it's not necessarily anything new because many of the concerns about the aggressive growth of chinese technology companies were already raised by obama and by his commerce department and it was actually them who started many of these proceedings against the zte the chinese firm that now has been brought almost to complete stop right because of the sanctions and because of the bans that we imposed on american companies that are dealing with it so america itself under obama was quite cognizant that its own technological hegemony was slipping away and we often do not pay attention to it but you know in fact america especially after the end of the cold war did enjoy almost two and a half decade of almost uncontested technological hegemony where much of the world depended essentially on their services whether it was google or whether it was yahoo whether it was microsoft and we did so without necessarily drawing the necessary political implications and conclusions from it right so for quite a lot of countries it seemed absolutely normal that you would go and introduce services from google or facebook amazon into your own national infrastructure including national security issues including you know health care and education and it seemed quite a reasonable thing to do because there were almost no political concerns at least about the provenance of those two systems right i think that that techno hegemony if you want to put it that way was thrown into something of a crisis mode as we started seeing revelations coming first from edward snowden with regards to you know the way in which american tech firms seem to be all too close to america's intelligence agencies we have seen another challenge to it more recently with regards to the cambridge analytica scandal but above all we have seen a challenge to it coming from a source that few of us in europe anticipated coming from china whereby now we have to essentially in europe make a decision between chinese suppliers of the same services and american ones because ultimately when it comes to oei and when it comes to artificial when it comes to cloud computing we still see american firms dominate they see right so both facebook and google make much of their money i think 53 and 56 respectively abroad they do not make it in america so most of their revenue comes from abroad in the case of china it's still relatively trivial so chinese firms mostly serve as a domestic market right but they start expanding and i can assure you that within the next five years the crucial decision facing unfortunately most municipalities cities and governments in europe would be not whether what kind of ai we should be using and developing it would be whether we should be buying from american suppliers which lead and feed our data directly to the national security service agency or the chinese ones which uh feed it directly to the chinese communist party i mean unfortunately that's the sad choice and dilemma that is looming on our horizon in europe and despite some of the more recent moves by european politicians who seem to have finally woken up so you know you see emmanuel macron making all these big pronouncements about the importance of ai what did he dedicate to it he dedicated out of the national budget if i remember correctly three billion euros to the development of ai you know that's what china spends in a week on developing ai so it's not i mean i'm overstretching it a bit but just look at one program so i think the one of the more important projects i don't think you mentioned it but the project the program for the development of ai which is their strategy until 2030. it's different from made in china 2025 so there is also another strategy which was announced last summer and it says plan to development of ai in china with the goal of achieving total world domination in ai by 2030 that commits if i'm not mistaken 135 billion to the development of ai just from the central government budget right so we're not talking about local governments regional governments we're not talking about venture capital funds of big tech firms in china and all the private investment that's coming to china for ai right uh so europe has a hard time catching up but it's also not obvious what europe wants to be doing in this field and i think that's one of the more problematic area because even the initiatives that were announced by macron they do not as far as i understand in any way seek to create an alternative poll you know an alternative system to the chinese and to the american ones primarily it's a matter of enabling american firms to set up european campuses in places like paris so that they can actually come and hire all the best talent in europe and put them to work in american firms which will need that talent in their global warfare ai with china right so europe at this point is a nice place to live with relatively uh well-trained talent but it's not anymore a driver of some kind of an alternative development for er which of course it could fulfill if it really wanted to think about this matter strategically and if it could politically and economically there is clearly a large gap in the market so to say because ultimately we probably want to have ai systems that do draw on a lot of data that we contribute but do it in a way that do not compromise our privacy which we actually in europe cherish still quite a bit which is not the case in china for cultural and political reasons and seems to be not the case in america anymore but we want to build it in a way where there is a more robust distribution of power between various actors involved in this so it's not just the central government as in the chinese model that gets to appropriate all the power about ai in its use which clearly it's a fantastic system for a policing as you know from china and b for making sure that people stick to the uh kind of script if you know what i mean so in china for example now how do you become a good credible citizen of course you take a loan like happens to be the case in many european american country so you take a loan and then that loan that constrains the social credit system that you've heard about is one way to make sure that everybody can qualify for a loan can take that loan and can actually then behave reasonably well to pay it back we do know that now they're using ai to actually analyze video applications for the loan and judge based on those video applications whether the person is credit worthy or not so essentially you sit in front of a computer you're being asked a couple of questions you record your video answers to it you upload the video to the bank and then the bank using ai analyzes your facial expressions uh as you respond to the videos and then makes a judgment as to whether you deserve a loan or not so for those kinds of users of course centralizing power and data and capacity in the hands of just one industry one firm and one government it can be fantastic i suppose that that's still not the world which we in europe would like to some extent to have and to build which means that you know we can have a system whereby data is collected locally to some extent um but not everything is done at the center level there are plenty of things that can be done by local municipalities and town halls and communes right in cities in a way that would actually be uh not as draconian as it would be if it was all done by the national police national intelligence agency or national security agency right and i think that there is nothing for us to be scared of that in a sense we can think of alternative ways to build transportation services that will be as flexible as uber but they will be entirely controlled by the citizens or by the municipality or by the municipality and the citizens of common and they will be driven by some kind of ai which clearly will be largely financed by the central government but it would nonetheless be administered in iran locally and i'll tell you why i think this is such an important question to resolve so one thing i didn't tell you about digital capitalism and the way it's going uh is that you see immense concentration of ownership of who owns these firms and what kind of money is poured into this firm so many governments and many sovereign wealth funds and many big institutional investors have recently discovered that so-called digital platforms and services of artificial intelligence are probably likely to be the most profitable investment they can make in the next decade given the paucity and the lack of any other investment option so given that everything else from bonds stocks will pay them in the good case in the best case scenario four or five percent many of them are more than happy to lift their stringent investment requirements and actually invest in startups where they would never invest before which explains why a company like uber is funded by essentially a bunch of sovereign wealth funds goldman sachs and a number of other institutional investors and that is the case for many other technology companies they go and take a lot of money from institutional investors and sovereign wealth funds and in many cases if you look especially at the gulf countries in south east asia those sovereign wealth funds borrow the money so they're very happy to borrow chipotle invest into companies like uber and airbnb harness all the returns and then be very happy now i'll tell you why thing is problematic so if you look at the biggest investor in the technology space right now probably a company that many of you have not even heard of so it's a japanese company called softbank so softbank is an immense firm that has its hands in a number of businesses they run a mobile operator in japan they own a mobile operator in the united states but they also own a fund called vision fund which is quite big it's around 100 billion dollars and they are rumored to be setting up another fund which will be 200 billion dollars so with this fund they have attracted a lot of institutional investors including by the way the sovereign wealth funds of saudi arabia of united arab emirates there is a sovereign development of bahrain that's discussing to join them and they're going around and buying all the companies that they can that do artificial intelligence and they do robotics and so if you have seen this viral video that was flying around the web a couple of months ago of a robot dog opening a door there was a very popular video that you know got a lot of traction online so that's a company boston dynamics that used to be owned by google and now it's owned by softbank and its primary contractor is actually the military and the pentagon so uh and softbank by the way it's not that it's just a company with very deep pockets it's a company that just knows how to borrow very well so the actual debt of softbank is something around 130 billion dollars it's a company that borrows very well and very cheaply and then with the money that they borrow and that they raise from other funds they go and purchase all the promising firms including by the way uber investing a lot of money in airbnb and so forth the reason why it's very important to understand is that you have to grasp something very basic once a company like uber takes over uh your local transportation service you know and i'm all for having a transportation service that will be much more automated much more flexible it will be actually much more plasma to use so there is no argument from me about that once a company like uber takes over that service you clearly see where things will be going five years from now and they where they would be going is in a very simple direction right now uber is losing money almost everywhere it operates but especially in north america they lose something like one billion a year in north america only right and the reason why uber can afford to lose so much money because uber is able to tap into the deep pockets of institutional investors and big funds like the one of softbank so they can afford to lose a lot of money while all of their competition essentially dies out and gets destroyed and then it's up to them what they will do later if they can they will raise the prices which will be make a lot of the customers unhappy or they can do something else they can essentially accelerate the process of automating the driving process which is what they have been investing in for the last five years initially replace the driver with the self-driving system which will allow them to keep the prices where they are but get rid of the labor entirely now once you get rid of the labor in that case entirely what you're left with is a company owned by the likes of softbank goldman sachs and the government of saudi arabia that essentially operates your transportation service in your local city on your local town which leaves very little money locally because there are no drivers who drive and get paid and buy services locally all of that money was maybe if somebody finds a creative way to tax uber some of that money might stay local as taxes but much of that money goes somewhere else and that i would argue is the generalized model on which many other services from healthcare to education to insurance to you know traffic optimization will be offered in the next decade and if that's the model on which many of the services will be offered you could not possibly imagine a better future than actually instead of surrendering all these services to the likes of softbank and uber and their friends in saudi arabia and elsewhere instead you can actually have a system where citizens and cities and municipalities have some control over it where they actually own it right and where some of the profits and revenues made with that system locally are returned back to the citizens themselves in part because a lot of the data that feeds into this automation artificial intelligence behind the system already comes from those citizens and that i would argue is a system on which europe can build a somewhat different model with the right configuration of political and economic factors i'm not arguing that that's going to be very easy to do because it will go against many of the even basic ideological premises uh off uh of the project in which europe actually is found but nonetheless i think that there is a niche uh and um i would argue by the way that will be exact same outcome if the role of uber is played by dd you know the chinese firm ultimately they share the same investors uber and didi both have solved bank as their main shareholder so at this point i mean we can go and find uh some interesting differences between national capitalisms and the way in which national capital reacts to incentives and pressures differently but ultimately at the level of ownership and who sats the direction in many cases it's the same global firms so on this i will probably be concluding just by saying that you know the the way in which artificial intelligence will restructure local and national economy to me seems quite clear it also seems obvious that we are living through immense consideration of power of doing it in the hands of just a couple of firms in the united states and in china i think there is still time provided that enough money will be found to invest into the development of such systems in europe in the absence of that i think having more centers of this rather than one meaning that having something very robust happening in china and not just in the united states is overall a good thing and on this i will end thank you um artificial intelligence 2.0 envisages documents uh an investment until 2030 in the past congress uh the communist party the president mentioned moderate prosperity program till 2014 i.e 100 years from the creation of the people's republic of china before entertaining questions in the public debate on uh ai in china the issues are connected with privacy the head of baidu who declared a national tv that chinese are available to lose their privacy if they have advantages a middle class is being developed there was a very fierce discussion on the environment in china something is considered as a hurdle to ai is the educational system which uh historically has been based on daokau uh an exam like the baccalaureate which opens the doors to the university or the labor market and it's a very rigid system as usually said artificial intelligence is in the curricula of schools and in a computer science university they also has used drone you know filming the students so that they can indeed get used to what they are studying in practice let me add that i agree with what you said something we call censorship is experienced in a different way by the chinese people it's a real barrier an economic safeguard to protect a local company's development facebook youtube and gmail what's up uh in a way censored but as a matter of fact the chinese people did not use them because they had other products such as run around like our myspace so there was an evolution of applications in china so not so that as you said in china people uh don't uh go around with wallets and purses they use which wechat which is a consumer to consumer app in china to book a visit at your doctor or book the restaurant and so on as a matter of fact in europe it works as a business to business app it's a company facilitating the entry of western companies in china are there any questions we collect some questions and then you will answer um thank you for the talk um i'm a student here in trento and uh i have one question uh we're in uh we can call it an impasse that of the single script as you call it that's real time censorship it's done there's a chinese drone maybe out there so uh we're in impasse and do you see any feasible way of distributing these centralized systems and what what kind of system do you imagine that brings us from this script to something that maybe your your friend clay shirky would call something like many narrations okay i would like i appreciate them and thank you also from myself for excellent talk and uh i'm fascinated about your proposal over a supposed poll of artificial intelligence in europe um at a level and we generated a level of municipality i think i i work in a municipality and i think that there is a um a huge uh this is the the biggest um amount of data that that you can um based on for doing artificial intelligence on but um i'm quite skeptical about the the critical mass you can afford to build in such a way i'm i appreciate a lot what barcelona is doing in this in this field and now using open source uh software using data from the citizen and so on and trying to build the services based on the on local owner data how do you think we can cope with it with this in the future in europe you are faith we are facing a level of non-coordination in europe about uh about all the other struggles uh we we have faced from the other part others hi thank you for the talk uh do you see this as a further threat moving from uh online to offline we have seen recently that amazon has bought wall foods and so this is first the first move it would be very interesting to to know what what you think about okay responders saying that we should seize the means of productions i love that implicit message in there um a question what's your take on gdpr is it too little too late or um you know general all right so regards to how we can i think some of them are kind of overlapping but so with regards to how we can deviate from this centralized system script right that you asked about um i think there is no way but to hope for a centrally that and enable decentralization i i i know it might sound paradoxical and to some extent but now but i just don't believe in the ability of baltimore players including by the municipalities of being able to match the capital investment that these firms are making right now into ai so if you're really talking about being able to replicate some of the functionality you really have to get real and understand the scale of money involved so despite you know everything that a lot of economists say about this current economy being dominated by interiors and nobody investing into anything if you look at tech firms they actually have very high capital expenditure so amazon spends i think more than any other company in the world in r d which is around 15 1 5 15 billion dollars a year google spends roughly 13 billion and facebook is somewhere around 12. i don't remember if i don't uh if i remember it correctly so it's a lot of money so if you really expect that you know the municipality of trento even if it joins forces municipality of bolzano will be able to somehow magically pull off that number out of uh you know head i think that's not going to happen so it requires but that can of course can be done differently you know you can do it in a way where you will find a niche and a role for cities and municipalities and you would be able to coordinate it with the nation level i mean in china they do it i mean they'll derive regional and of course their cities are much bigger and much stronger and have much more money but i mean there is quite a lot of activity happening locally which you do not really see happening in europe but that's sad i think i understand the question about the practicality of all of this given the current mass in europe in italy and in many other like nation states i think the only possible solution here is to put this question on the political agenda in a way that now i'm not talking about in a way that all political movements and parties will have to confront it as a political issue and not just talk about taxation and not just i'm not saying that taxation is not important i'm not saying that regulating the big economy with proper labor laws is not important but that crowds out almost all the other high level discussions that need to be had including on issues like industrial strategy and what does it mean to have an industrial strategy when it comes to artificial intelligence and what it means to have industrial strategy when it comes to data right and what kind of ownership systems you need to have in place to enable that in a much more decentralized manner and the reality is that very few political forces anyway in europe that i see i prepared to tackle this question as a political question because you cannot just because if you really are serious about it you also have to problematize it as a european issue it's not just the failure of nation states it's a complete failure of the european union as a project to essentially muster up and gather up enough industrial strength and to actually reaffirm the fact that well in china industrial strategy that's actively preceded by the government seems to work and in europe its absence does not seem to work and i think it's a very basic recognition that anybody who looks at the data honestly will be forced to make and the problem of course here you will need to be able to pursue a critique of the european project and the reaffirmation of some kind of european sovereignty if not national one in a way that would not make you sound particularly xenophobic uh nationally and it's just a problem that most leftist forces in europe that's still around and are not fake left forces have not managed to essentially resolve so and it's not a problem that you can resolve just by understanding the role that softbank plays in contemporary digital cabinets so i can tell you more on that and we'll need to go far beyond the problems pertaining to the digital field and just two last questions offline online and how much of it is moving offline you know what i try to to do today is to kind of hint at some of the dynamics that drive these firms and you know i think i'll be there by now and say that what drives these firms is they search for profitability and in that sense they don't care whether they do their activities online offline they they pursue the activities where they expect to generate most profit so if amazon expects to be able to get something from the offline data if it will allow it to offer better web services in the online world they'll do it so in that sense i think that this firm's a very uh kind of internet agnostic you know for that a cyberspace agnostic you want to put it that way then ultimately they don't care as long as they can put sensors everywhere they'll take all data you have offline data online data you know your thoughts your dreams you know they have a smart mattress they will extract whatever you dream about at night they don't care you know as long as it contributes to the bottom line they'll do it so i think we have to be also much less rigid in trying to imagine this industry as somehow operating in cyberspace and there is a another industry that's not i think ultimately it's that very conception that caused so many problems in our ability to understand what this firm's about with regards to gdpr i mean i would really love to say a lot of nice things about it but you'll read it all by email no uh well i mean look gdpr i think has it's a legal instrument it's a legal instrument which for certain types of tasks i think can work remarkably well uh and you know i think it also has a somewhat different um idea about the necessity of common ownership of data that i have so there are certain things in my vision that would be very hard to implement within gdpr but ultimately i wouldn't look upon a legal instrument to be able to articulate an industrial policy for europe or any of the individual nation states right ultimately that's what you know you can say whether it's about seizing means of production or not but ultimately that what i think is missing and you cannot possibly let european politicians and national ones get away with saying that gdpr will solve the kind of problems which it constitutionally by default is incapable of solving so being conscious of these things that gdpr can solve you know i welcome it but i'm conscious of its limit and i think the danger is that it will mislead a lot of european politicians into a sense of comfort in which they've been for the last 15 years already thinking that they actually have a grasp of the situation and they have to means to resolve it while in fact they don't you know my main fear is that europe will find itself in a position where on paper theoretically privacy will be protected with all the possible legal instruments imaginable but in reality all the other economic rights and social rights will in some sense become meaningless because everything from education to health to transportation has been overtaken by commercial providers of those services so in a sense you can be saying that you have a meaningful right to the city but when your city is run by cisco ibm i mean what does it actually mean right so google by the way which is building its own smart city programs you know sidewalk laps so so that's my problem so you know i i think regardless of the cambridge analytica scandal and everything else i think the discussion about privacy in europe is quite robust but it's not very robust about all the other issues that i think should be up for discussion and that's what bothers me so i'm all for gdpr you think it's a necessary but not sufficient conditions thank you okay anthony