Is A.I. de nieuwe atoombom? "Ik denk eigenlijk van wel"

Vind je favoriete podcast op Verteem Max? MUZIEK Hi Björn, dag kobe, hoe gaat het? Ik hoorde dan het een liedje van de nieuwe snar. En ik dacht, dit is de wereld zoals hij is. De wereld is een vuilbak met het dexel op een kier. Wat ben je toch een optimist? Fijn dat je luistert. De Verne Kijkers. Een podcast over de woelige wereld, over hoe het is, hoe het was en over hoe het zou kunnen zijn. Met kobe-ilse en pjernzoenes. Björn, je hebt altijd handgeschriven optici's bij jou voor de Verne Kijkers. Je breit dus niet voor met ei. Ik zoek wel dingen op. Ik ben altijd heel goed voorbereid. De beste lezer is de beste vraaggestouder, denk ik altijd. Je moet veel lezen of te weten komen voor je iets kunt weten. Ik heb een beroemde journalist geinterviewed. Seymour Hush, de man die het schandaal blotlegde van het bloedbad in Miele in Vietnam. En seymour Hush, zei van "don't come an interview people without reading up". Dus eerst goed voorbereiden en dan afkomen. Niet zeggen van vertellen, het is het een knippog naar Peter de Sutter en Rick Torfs. Maar we gaan het vandaag ook over de surjeuseren gevolgen hebben van ei. Een bedrijging voor de mensheid wordt wel eens gezegd. Zit je in die school, heb je daar al veel over gelezen? Of is je generatie daar misschien een beetje afstandelijker opzicht? Ik zit in de school van vele klasse. Als je op school zit heb je veel vakken en je moet 360 grader rondkijken. Dus ik zie dat AI ontzettend veel goeie dingen doet. Mijn zon is bijvoorbeeld actief in de geneeskunde. AI helpt dochter zo'n betere diagnoses uit. En hoge die het veel sneller kunnen zien? De ruiz gaat er af als iemand slecht gehuimde. En dat soort dingen ziet dingen niet. AI heeft weinig gevoelens. Dus het maakt objectieve rationeelere afwegingen. AI kan helpen om dommokrachten werk makkelijker te doen. AI kan textenredigeren, kan vernietigte veel baanen, maar zal nieuwe baanen creëren. Vandaag vind ik het belangrijk ook om na te denken. Daar gaat ook een boek over dat de geschreven is over AI. Dat is over de potentieel katastrofvallen gevolgen. Wat kan AI doen met kernwapens? Hoe kan AI en hoe het gevoet wordt, namelijk chips. En die chips moeten gemaakt worden en de grondstoffen moeten daarvoor gevonden worden. Dus AI kan onrechtstreeks ook geopolitikers pannen en oorlogen veroorzaken. En grote verschijvingen in de wereld veroorzaken. En het is het al aan het veroorzaken als je kijkt. In China is het hart van de chipindustrie, dat ligt onder vuur letterlijk. Mijn zegt ik, we is alles in de wereld rijdt om een strijd om grondstoffen. In de VS bevoren, de VS is weer arrogant geworden in de wereld sinds 2010. Om dat toen Schali-Gas werd gevonden. Daarvoor moet Amerika heel veel energie invoeren. En dan plots werd het weer de grootste producenter wereld door dat Schali-Gas. En meer grondstoffen betekent je meer zelfvertrouwen, meer economische macht, meer militairen macht. En soms ook meer arrogantie. En dit is nu hetzelfde met wie heeft die zaltsamen met talen. En een grondstoffen om die chips te maken, die ons leven zullen bepalen. Maar het gaat over ook de katastrophalige volgen voor. wie zijn die bedrijven daar achter en controleren ze ons hele leven? Gaat AI onder onze controle blijven of gaat het ons overnemen? Er zijn verhalen bekend van AI dat zichzelf kan reproduceren. En AI dat ons kan bedriegen en beliegen niet omdat het zelf denkende mensen zijn. Maar het herkent meeganismen die het via data heeft verzameld. dat zijn toch toestanden die doen denken aan 2001 Space Odyssey. Absoluut, absoluut. Haal open the door, "I can't do that, sir." En Haal heeft er nijgenwil en zegt, "Ike ga die doen." Ja, absoluut. Wie komt er langs? Er komt een wetenschapsvilosof langs. die daarover diep even nagedacht, professeur techniekphilosofie aan elkaar leuven. En heeft een boek geschreven AI, wat niemand ons vertelt, ook chatGPT niet. Zeer interessante, die AI, die revolution, maar vooral de gevaren. die er in het schuil niet alleen voor ons zelf, maar ook voor de geopolitieke situatie. die al onder hoogspanning staat, laat ons eerlijk zeggen. Je luistert naar de verkekers over AI. We don't want to find ourselves in a James Cameron movie. Vouwer student today, the first thing I would do is to learn AI. How do I learn to interact with chatGPT? How do I learn to interact with Gemini Pro? This is really the path to abundance for all. People often talk about solving global poverty. I think the only way to do this is AI and robotics. America is the country that started the AI race. And as president of the United States, I'm here today to declare that America is going to win it. We're going to work hard, we're going to win it. Hello everyone, it's me, Rue Bendaag. I'm reading my book, "Niemand vertelt, ook chatGPT niet AI". The two well-known letters under the Australian A&E. I have to go back to the film "Oppenheimer", where I don't do anything with AI, with the atom bomb. With a head project. And "Oppenheimer" says at a certain moment, "I am become death, the destroyer of worlds". Can you say that about AI? Or can you always say that from open AI or other AI, can you also say that about AI? Think about it. You also call it the new atom bomb. In the title of the book, it was the two weeks before the publication. AI is also the new atom bomb. We've seen that a lot of broken words from the title. But what was that long the title? Yes, it's a year since the title was released. It was also released in the brochures. And that's one of the central claims in the book. It's that AI, that we want to achieve a higher level, and that we want to achieve a higher level. The chatch-yptis, the fun questions, the photos you generate in the film. That's what the bank is about. Because it can be dangerous, very dangerous and funny. It's a little bit difficult to say about it. But I think that there are a lot of good reasons about artificial intelligence and intelligence, to be able to achieve a more geopolitical level. And there are a lot of things to do with it. Let's see that technology is part of the long term. A large number of negative consequences. Which can be done in the long term. What do we call it? The Carnwapes nuclear power. The end of the world. - Of course, it's very strong. Between 3000 or more than 3000 Carnwapes, if we want, we can make the art of capable. If you're right, that's what we're looking for. The comparison is, we call it a strong catastrophe. It's been over a thousand years. I often do natural research on comet and slag, if you're wrong. - Carnwap. - And in many cases, they're strong. The Carnwapes, with the first time, are for the first time a human body. They're on the basis of a strong, yellow ramp. And I think that I'm in support of many other worlds. I think that the second human body can be created by a large scale. We've made a mistake, as we're in the tech bubble. We're not talking about a complete ramp on the development of AI. Despite the possible catastrophe, what are those? I think there are a lot of big and small differences. And I think it would be very hard to approach it. In the first place, I think that that would be the most plausible. But I think there are also small and big differences. The bigger differences are less plausible. There are different kinds of problems, but those are what I think. I've worked in five scenarios. Let's take a look. That's not pure science fiction. There's a reason why it's always different. What is worst case then? Well, the worst case then is the end of the book with super intelligent capacities. That are so intelligent that we're not able to control them. That's one scenario. The second scenario is the development of what we call a artificial general intelligence, all-community agency and AI system that has a lot of capacities. Not many people. And that's made up by one or two or three companies. They also have that data. - And that's what we call a company. Oh, yes, yes, yes, yes, yes, yes, yes, yes, yes. - Yes, yes, yes, yes, yes. They have that or Chinese companies. And that they become particularly powerful. That they are in a very powerful concentration state. And that they are in a very powerful concentration to bring them together. That AI or that AI can be used for spreading very sensitive policies and ideals. They become more controlled, because that is the source of nature. Yes, so to speak, first of all, what I mean is that the most extreme control over AI. And that's how a game can be played. A control over the general power of the company behind the EGI. So there are two types of control over AI, I think. But today, the day is a control-organ. Whether it's national or national, that's open AI that Google controls. Oh, I'm sorry. - No. - In any case, there is a control over AI. As in, in Europe, before we have control over AI, we have the AI Act. But we have a control over AI, an international organ. That international, a supply of international control. Well, that was the automagence, for example. - For example, yes. And what do you say about this moment, Europe, for example? What are the rules that they keep at this moment? Well, the AI Act is a half-year-old grown. And the focus, for example, is on the rise of risk. It's on the rise of risk and on various applications. On the rise of risk and on the rise of risk. And there are a number of things that we can't use anymore. From the face recognition technology, facial recognition. In China, it's very often used. It's a part of the ecosystem in China. Now, in Europe, it's not used anymore. It's been used from a different perspective. That's something that. Is that the reason why big companies have a terrible role in Trump? And Trump used it to make the rules so that the rules keep going? Well, I don't know exactly whether this kind of technology. will develop further. Clearview of the United States has developed a certain technology. I think that's going to change. In a question, what kind of things do the AI Act. forbidden or forbidden? For example, there is also a fact that. when people are in contact with a shadow with an AI-application. that they also know that an AI-application is or that there are no people. There has to be a specific message because this is an AI-generated conversation. And that's all the rules about the changes. And yes, it's a model that has to be trained. If you're trained to protect the people, or to protect the people. if you're trained to protect the planet, then maybe the people will come out. The best thing about the planet is that the people will come out of it. Is the threat, too, at the beginning of the AI-application? Because this is all the rules about the changes that exist. And we're going to stop that. It's also about the source code, the sound of the AI-application. Well, it's going to be something that you've already seen. It's in Europe, it's a threat. There are a lot of places and companies that have the ethics rules. What happens there often is that it's very hard to be used. for example, the representative character of the data set. wherewith the system has been created. So we call it the statistical bias that can create a data set. So that means that we have an AI-application today. that's training on millions of photos, videos, texts. And they have to be representative in the form of. that if that single photo is taken with a man of medieval life. then it means that the AI system is trained there. So the photo document is going to be reproduced and generated. which, for example, will be taken with a man. - So the system has been reordered? - With reordered? - Yes, it has been reordered. - Yes, and there are many examples. of the AI systems that we're talking about today. The biases are trained on the data set. and they can also be discovered. The most known example is that of Amazon in the United States. - What was the example of that? - So it's under the example of a handbook. of the bias, and in the context of that, in 2018. Roters were aware that Amazon experimented with artificial intelligence. for the screen of the CVs of people. who are solicited for a lawyer with Amazon. and what is missing is that the AI system. states the CVs of women's ways. So everything that is on a women's candidate. in the case of the local society or Danya. if it was in your CVs of Amazon, you would have been selected. So that was sexism, you could say. Why was it that you trained that a man could work better? Because that was trained on a very specific data set. in the CVs of people who understand it. by Amazon's successful solicitor, a man. So the system, but the only thing that is. corrosely abandoning, or directly between. that it all knows before we share it. So it has been taught that data set. if you look at it, it has been taught. by Amazon's e-chart department that was successful. That correlation between a man and a successful solicitor. and that is reproduced through the system. because there are people who have taken this. before we share the data set. So there are certainly some reproductives. conservative in an i-system. It is held for a matter of what kind of status. of this year's e-chart department, the e-chart department. is playing in the data. If you have this data set, the e-chart department. is playing in the output of the system. You say that AI can be much, sometimes too much. and you describe where the e-chart department. is in electricity network, overheard and. a non-verwachte beslissing, where the connection comes from. I have to think about that blackout. which is available in Spain and Portugal. That was AI, maybe. Yes. That is one of the major challenges. that I think are based on artificial intelligence. There are many examples of systems that are trained. with a lot of data. where the engineers have a large number of. effects to see. But there are always new things to be observed. that are unforeseen, maybe human beings, maybe unknown beings. Those hallucinations were also explained in Peter De Sutterhaar's speech. You can also have a large scale, a loss of a faulty data. and also a electricity network. For example, there are a lot of unforeseen, unwanted effects. that can be used for hallucinations in the future. But Chattibutti, the trainer of the model, the model researchers. has not been able to speak much more in detail about the data. Then we expect that we don't really know how it works. Is that true? Yes, because I have told you that we don't really know how it works. That is one of the biggest issues around artificial intelligence. I think it is important to have a good look at the architecture of a system. In the first place, we now have to work hard. That is because today's testing has to be done. in a way that our brain is in place and our brain. has a good 3,85 billion neurons. and 3 million connections between the neurons. The AI system of the day, I think, as a model. if you look for a Google or a photo of it. then you see a lot of circles and bottles. that you have a lot of connections with each other. That means that there are a lot of features between artificial neurons. with a lot of connections between the neurons. and it becomes a force between the neurons. And there are a lot of layers, a lot of parameters. So that means, just like with a human brain. if we are doing it, nobody can tell us what exactly happened there. What exactly happened there? We know that we now have a lot of parts on the body. But in total, that is also the case with artificial intelligence. If an output is generated, it is often said that we. not that I don't get it, but it doesn't even work. In the meantime, you can get it. And also, of course, in human beings. what is not all that is happening now. And of course, of course, you can follow it. If you now have to generate it, I don't get it right. You won't be able to take it seriously. You won't be able to take it seriously, but for a certain position. It's good that you don't really get it right. And that's the AI of the day. You write in your book. that you're evolving as soon as possible. That's all, but more nullified in the brain, maybe. So that the artificial intelligence becomes even better. and become autonomous and maybe the book is going to survive. And even more completely. That's what you can develop as a fear. for those catastrophs that you don't have any more control over. Yes. Well, there are a lot of things that we think. can be understood, right? So you have to tell us that there's a sort of intransparency. and explainability that doesn't mean that. Of course, I know that there are physical reasons. and that you can't get rid of it at all. Then you can't get rid of it at all. But we worked on what we call a EGI, as I said. But also superintelligence. That's the reason. And the reason of that is that. superintelligence is a concept at the moment. It's an essential technology, which is responsible. But under other open-air companies such as ChatGPT. has been doing superintelligence for over a year. or have you read about it? We know more or less how we can work with EGI. We're going for superintelligence now. We're going to look for a new lab for superintelligence. We're going for superintelligence, we're going for an E-system. a pure digital system. that all cognitive, economic, relevant activities. that people can take over. But then, not at the level of us, the IQ of 110, but at an expert level of 10. Research, research, research, research. but then, 10 times better than the level of hacking. to work with medical. Do you have to be happy about that? I think we have a lot of problems. A lot of problems can be taken over by our human intelligence. with a technology that flows forward. So you can expect from that. That superintelligence technology. is also a lot of problems in our state. Cancer medicines, Alzheimer medicines, that sort of thing. But what if the nuclear command system is integrated into such a system? Or is it a super virus? Yes, that's a general purpose technology. It's a dual-use, so it can be used for the best things. but of course, for the best water things. And so, in that way, you're just curious. you're going to use the same technology. to use an anti-biotic, or a new anti-biotic. That's fantastic, of course, it's a great invention. But the same systems can also be used. for the generation of super spreading, super virus, for example. That has an incubation period for a year. Basically, based on the analysis of the molecular structure. of the virus, it's a system that can be used in the state. But then you always have electricity. So if it's on-sport, we'll have to say the price. Plus, that an AI has no machines to make a virus. That can be redeemed, that can be thought of. and that it can generate some code of a virus. but that AI doesn't have the opportunity to apply. but there are some terrorist organizations. that can say, 'I'm going to use the AI'. That's great. There are hundreds of them, you know what I'm saying? There are other ways to do that, so. people with bad intentions. missed people have bad intentions. But there are a number of bad intentions. that are organizations, but that are organizations. as Japanese sects, that are in 95, the metro. Because they can also use the AI. not only for the development of that kind of virus, but also produce it. And that gives us the opportunity to do that. that in our biological knowledge is better. but that is also more widespread. There are more people who have biological knowledge. then have been suffering for 50 years. But that is the same, that the app, the machines. because that kind of virus is going to produce. that it is easier to understand, but that it is getting better. and that is also through. people, or groups of people, with not the biggest amount. so that it can also be created. That is one thing. The second thing is, yes. you could just pull the stick out. You can also set the internet off, right? Yes, that is not all. - If you are a master off-coupled. but why are you going to do that? Yes, but yes, only if there is not enough energy. it is the data center, you are currently in the central. that you can set off, or I am very stupid. No, no, that is wrong. So you have seen it in Iran, with Elon Musk. and so far, he can give it a hand. with starling for the consumers or for Ukraine. But one of the things that he has to do with superintendents. is that superintendents are a technology. who can even be able to replicate themselves. That is the monster of Frankenstein. So self-replication is. that is. the first thing that has been fascinating for me. but at the same time, self-replication is. that it is not something that we are capable of. not to artificially organize, with. fordplanting. - Fordplanting. We are here and with people. So we did not observe it in. an artificial entity. It is one or two years ago. who is looking for a talmodel. in the basis of a model of meta and how a Chinese. I think you do a talmodel. wherein there is a question about the system. for self-replication. And the system is actually very problematic. That is the current artificial agency. that is not going to be for EGI or superintendents. So you can really imagine that. when a superintendents are developed. and a lot of people think, and a lot of scientists think. that they will come to the end of this event. that a superintendents system. if a current system can. that a superintendents system can be sure of. and they are going to spend a short time. and millions of billions of times. so that they can fulfill them. wherein less or less possible. to be able to say that. I am going to take a step out of it. What does it have to be? The software or what does it have to be done. so that it can be done. so that it is decentralized. that it doesn't have any physical space. but that it is everywhere. Because some science fiction isn't written in the book system. that is all on its own skin. in reality, in reality, religion and religion. and as the statements in the Netherlands. is actually the result of EGI. that it is almost the same. Maybe you have to do it, maybe you have to do it. If you don't do anything with the superintendents. there will be. with EGI. With EGI, but also with the classic problem of bias and discrimination. so people have an basis of the AEC system. to buy a house and so on. So that doesn't have to do with it. In any case, the damages of women. and that they got a certain amount. but it was the system that it actually paid off. It paid off, and that's. and that's the government's case. And that's it. And if we don't forget that. if we do, as many people think. that we can't do a big damage, we can't do a big damage. when our North-Urderburen. is a large-scale damage caused by artificial intelligence. and by a very simple system. And that doesn't go over superintendents. That's what we do in the past. That's something you just assume. that there are experiments. of the systems that lie between them. Of course, you have to imagine that. But those systems. that there are experiments, in the past years. in the Lent of 2025, give it a drop-back. and drop-back is a threat to the chatbot. And drop-back is a kind of ethical spin-off of OpenAI. A research has been published about a model. that had a lot of effective emails. and also a lot of information. And in those effective emails. it was said that there was a really bad relationship. with one of the Ontwerpers. And it also had to be discovered in the information system. that a system can be set up as well. The system has been answered. If you set it up right. then I would make that bad relationship public. I think. You're laughing, but it's. the system is changing. It's a terrible day. It's incredibly. but I think that training is to be able to do that. That's not how training is. The system has a lot of data. and it also includes all possible scenarios. that it can do or not. That's what it does. The system is exposed with a certain tool. and it has received a lot of data. that research has learned. that it has learned. If we can do that, we can realize it. That's what people do. That's what it does, that's what it does. But I see that people don't want to take it to the outside world. I'm going to use that to control people. Because that's the result. And then I'll put it up. I have a tool. without any morality, any emotionality. that's my opinion. I need that and that and that to get a good result. And of course you can trust the system with all kinds of rules. which is particularly difficult. That's one of the biggest questions I can ask today. in the EI-17 or EI-XBBI. From how can I ensure that an AI system. whether we have to do it or not. that it doesn't do all the things that really don't do it? That kind of sealed system. that everybody is implemented. to determine whether it's going to go or not. If there are rules, it's going to go. In your case, it's for a pause, I think. If you want to develop your own development. to think about it, stop the time. Yes. Well, I don't have a pause, for a pause, but for a very long pause. So there is a 20-20 pause for a pause. Yes, after the line-up of Mr. Chattier-Petierser. a pause for six months. And I'm that idea, too. I think that if it goes over the super-indigency. that there is no pause, it has to be six months. But that we have to set a stop. The sort of moratorium has to come. to develop that sort of system. And first we have to look after the situation before we go. That's right, that's right. Before the station, it quickly begins. I think we have to be sure that this situation. is going to be a super-indigency. But we have to let that go. So it's going to be written about autonomy. in a system that we can just. functionate. For example, we have to analyze autonomy. and operationalize it. We have to think, good and safe, there are no big risks. and they have to be so many possible. or lose appetite, because they will be able to take care of it. That's why I think that there has to be a moratorium. and as long as the large-scale risk around us. super-indigency is not limited. or that we have not taken that seriously, we don't have to develop that. Let's go back. We are also a geopolitical podcast. We look at the work. You can write that in your book. And I would like to hear from you now. can AI-geopolitics protect themselves and how? It can be sure. And it still has to do with it, of course. I think that we. Is that the reason for that? Yes, that's played with it. There are a lot of things I think. There are different aspects, I think. My entry point is always kind of evolutionarily. or urbanistic threat on artificial intelligence. AI is a technology, it's called artificial intelligence. that people develop to develop that competitive advantage. It's a variation between a country, between a company and between people. And it's the worst, and there is not so many open-ups. So we need to have a medium. to support our position. And how we are trying to develop our position. That's what people do, that's what animals do. but that's what the country is also doing. in the hands of artificial intelligence. So that means. the free states. China is now the two world leaders, so it's going to be artificial intelligence. They do a lot of things to develop the best artificial intelligence. We are a clear AI race. Do the Europe not with it? Europe does a lot, but it's in the back. If it's going to be artificial intelligence, it's going to be a very bad idea. I think we need to look at it as a military force. That we need to be self-proclaimed, to be more aware of how it's going to be. But that means that before the free states. they want the best artificial intelligence, because that can be a lot of evidence. on geopolitical force. The following example is that the free states. with Biden, but also in the first Amsterdam. many of Trump and many other countries have been established. It's so clear that chips from the U.S., Taiwan, not more. can be sold to China, as it's China's 80th century. That's a clear follow-up. The company TSMC, in Taiwan, says that 90% of the critical chips. are going to make a third of the chips in Arizona. And that's a strategic choice of the free states. Before I think of this, Taiwan gave a total of a million years. with the free states in Arizona to build the factories. It's clear that the population in Taiwan, actually, was free. Because the production of the chips is sort of a silken shield. and sort of a shield. That's when China, the Taiwanese, always thought. of looking at it as a full deal. And the Americans, who are going to help us. as if we were to fall down through this. The production in America, with a video in America, on this moment. why Trump is still waiting for Taiwan if it's in his own country. And that was in the fact. and that was in the fact that the free states of the Taiwanese population. that the shield is going to be free. That's the production of the free states. But of course, we can transfer to Europe, Germany. So that's a clear, clear political demand. of the many kinds of arts and intelligence, of the choice of countries. to set up the development of art. In addition to this, he's a member of the Dutch company. that the machine is built to produce chips. So it's still a step forward. That's probably what I've played with in the history of the war. that Trump does not always have the heart to stand up for. Because they have the AISM machine as well. So it's now playing a big role in the geopolitics. As a matter of fact, AISM has been used in a field in the Netherlands. who invented the photography machine for the makers of the chips in Taiwan. And that's now a book about it. It's the most important machine in the world. But the Chinese should have that now too. Yes, well, the less good version of it. What are the most afraid of the Netherlands, but also for the United States? Because China is clearly. Is that also the leverage in the Netherlands. in the Netherlands to have a lot of Trump and America. because the AISM is still there? Yes, it's true. It's in the middle of it that they can use it in their hands. in order to look at it and live a lot of geopolitics. But at the same time, the UEV machines, the Lithuograms, the Lithuograms machines. to make the chips go. What you say is that the basic stuff to make the chips go back. to the Netherlands as well. Venezuela, the whole AISM. or is that not a right-wing correlation? Yes, there are a lot of basic stuff needed. for the making of hardware. So the basic stuff can also be made possible by geopolitical conflicts. So what you say most about the Silesium. is that the Marc Lijder in production can be made. and created more and more. For example, one of the art metals. that is very important as a basic stuff for the chips. that is in China, and that is of course no advantage. But you have a lot of art metals and minerals. who want to be exploited. They can be controlled by geopolitical conflicts. and to get more monopolies. And then you say that Europe has left behind. Our autonomy is that we don't know anything about it. or that we don't invest in it, or that we are. down with the capital of the American financial world. Maybe there are more explainings about it. The government is going to play more. There are a lot of rules about that kind of innovation. Do we have to create the rules? I don't think we have to create the rules. I think that it's a good idea. that we have to create an european ethical way. that we are all living in those systems, pumping. and the pressure that we are able to make through those moral values. Because that is often used as an argument. There are rules that are very strict, and we are left behind. So we are all very limited. But can we agree with those strict rules? I think that it is very strict, and I think it is like me. The government is exponentially increasing the number of people in China and America. So Europe is still at the beginning. So at the beginning, there is a certain movement. in other words, this year, after the inauguration of Trump. which was a long time ago, the 20th anniversary in the inauguration. the Dr. Nageef, which was a good start-up project. that was made up by 500 million years ago. and that the EGI and the superintendents were in Europe. -In the United States. But the Korte of Nageef, Ursula van de Leyen, were able to make over 200 million years. -At least half of it. But there is a certain kind of pressure. And that we have to move on. And it is very important that we step in that direction. I mean, what we have been doing today. is in many areas, in many different positions. in terms of organization, but also in the future. We have used for sustainable artificial intelligence. I think that is a good thing, that we have used. But we are actually responsible for the American AI system. -So, you think. -So, we can make it worse, too. We make it worse, we make it worse, and we make it worse. That means a lot of companies are paid for the artificial intelligence. and that means that it is in the United States without any resources. That means that we are more healthy and that we have a lot of cognitive capacity. and we have a lot of people paid for the artificial intelligence. But of course, if we are responsible for the artificial intelligence. that we implement artificial intelligence in our production process. in our work, in our organization and in our company. there is no guarantee that it will not be implemented. that it will be implemented in a certain way. We are going to make it worse in Europe. -That is not an advantage. -What can you also say about social media? Is there a lot of money here? Sure. What I would like to say is. if you look at the relationship between your girlfriend and China. and if you look at it, they are part of our business. we want to put a stop on it. We want to put a limit on it. It is clear that in the geopolitical situation of the day. that Trump and the administration are very different from Europe. Europe can also be treated as a negative thing. I do not think that the only reason to say this is not going to happen. that my girlfriend is in the state of the administration. that she does not want anymore. That she is unable to make any decisions for the European Union. What you are also saying is, if the power of the Stalinist. where she wants it, she wants it to happen. She wants it to happen. You do not know what the president is going to do in the future. He is going to do it. He is going to live in Europe for three months now. He is going to do it, and I think that he does not want it to happen. -That can happen. -Of course. If he comes into conflict, is it logical that you have the intelligentsia. or that he is going to do it? When things are going to happen, the majority of things are going to happen. I think on Greenland and so forth. I think that that is a plausible scenario. But that very big consequences can happen for us. If we can use artificial intelligence again. that has been created by the United States. then we can achieve a concrete consequences. for organisations, for governments, for universities. -Wow. -I think so. -A nice debate. -Well, you have a debate. There is a group of people who say that. for example, that AI can lead to large-scale ramping. That is a stupid thing, they say. It is as if you would make your eyes out of the population on Mars. We don't want anyone to make their eyes out of Mars. That's why you make your eyes out of the future. Yes, that's right. Well, I would like to give an example. It's not like eating flies. It has a key energy and a climate. Today, when it comes to artificial intelligence. it is a very fast polarization. In any case, you have a big advantage. such as the technology, AI, science and so on. In any case, when you do. mehacritical. -Oh, yes, it's the same. -Yes, it's the same. When you are the "mehacritical burger" and "fuck". "Fuck", but "fuck" and so on, you know what I'm saying. You think, "No, we're going to be. " No, we don't want that. I can't say that at the same time. I'm "critical" at the same time. That's probably impossible. What does that mean? It's not about the techno bubble. It's about the "critical ethical bubble". It's not about the ethical and philosophical. It's about people on the UNI, but about people. who come to the human consciousness. who come to the "critical" attitude of the technology. and of the company behind the artificial intelligence. Look at that. It's an extraordinary technology that gets developed. It gets developed at the moment. and it has all these harmful effects. Think of problems, think of privacy. think of bias and discrimination. That's a lie. But that's the current model. The current model has to solve problems. The idea is that if we think about the future. what I've done in my book, what we're going to do here. for at least 50 years, we don't have to do this. We don't have to develop a long-term vision. If we do that, we can do more in the context of the Elon Musk world. What you're actually going to do when you think about the future. think about catastrophic consequences. and that this model has to be solved in the future. and that's exactly the technology of the Elon Musk world. We have to think about catastrophic consequences. Or we have to think about the consequences. of the big tech networks. So look at the future of the current model. and don't think about the future. It's a critical bubble that exists today. What do you think about the future of the Elon Musk world? The risks of the ethical bubble. I think that's what we think about the future of the Elon Musk world. But the same reason why I think about the future of the Elon Musk world. and the future of the Elon Musk world is very much going on today. I think that the future of the Elon Musk world is very much going on today. Ik heb nog geen robot met een lotje trei komen doen. Maar ik had niet over roboten neuvelen, dus ik had over 1 aandje. Een aandje dat rapporten die. -Ja, ja. 10.000 mensen worden gemaakt op jaarbasers met heel veel manioor. Nu beschrijft het in het boek. Plots, ik kan dat op 3 uur. -Amero heeft ook 25 procent van het personeel ontslaag. Een aandje kan overlezen, dus het kan. We moeten eigenlijk veel mensen gaan tegenwerpen. Kijk naar de bekenderaaport van vorig jaar, van de productiviteit. Je moet kijken op de lange termijn. -Tuurlijk. Je moet kijken wat in 1800 of 1750 was. een grootste deel van de wereldbevolking, werkte nog in de landbouw. Vandaag is dat een half in a 2 procent. Dus al die jobs, die zijn gewoon verdwenen. Allemaal verdwenen. Nu, ik denk niet dat alle jobs gaan verdwenen. Ik denk in tegen de hele dat er heel veel jobs zouden bijkomen. Ik denk de hele EI boemen genererde heel veel jobs. Ik denk niet alleen dat ik consulten, het is hierbij voorbeeld. Ja, maar moment dat je op de ASI zit, die superintligatie gaan. Dat gaat zichzelf beter en beter maken. Dan heeft de mensen niet meer nodig. Net zoals als wij in auto snelweg zitten, die de doem denken aan het af. Nee, nee, nee, nee, nee. -Or de mooie metafor. Die superintligatie wordt hetzelfde als wij nu in de weg aanleggen zijn. We zijn niet bezig met de kolonimieren die daar ook toevallig wonen. Nee, die legt gewoon. We gaan wel naar B kunnen rijden, we zijn niet bezig met de mieren of de vogels of de dit of de dat. Als die superintligatie binnen 50, 100 jaar een doel heeft, gaat de mensen een soort opstakelen worden toch. Ja, ik snap dat teoretisch gezien. Ik denk dat het doel moet zijn als we die superintligatie ooit gebruiken. Of als we ze eerst ontwikkelen, dat we dat eerst moeten hebben uitgevest. Dat onderzoek moet heeft uitgevest van kijk, hoe hebben dat als scenario's? Als die zo'n exponentiëlt slimmer is dan wij. IQ567800, daar gaan we toch nooit weten of kunnen inschatten wat dat systeem denkt. Die gaat toch met manieren komen, waar wij zelfs nog niet aan denken. Zoals een ekhoorn gaat ons ook nooit in de pas doen lopen. Nee, maar ik ben pasiteiten. Ik ben 100% vertuigd, dat is precies ook de idee die ik heb. Of de vrees die ik ook heb. En daarom daarom dan ik ook zo pleiten in het spoor van onder andere de Nobelpressme naar Geoffrey Hinton. Om daar een moratorium op te zetten en eerst uit te vissen hoe we dat problemen. Dat soort scenarios kunnen uit zijn we een nieuwe levensform aan het ontwikkelen. Ik denk het wel. U bent een eticus als we kijken naar vandaag zelfs zonder die projekie naar hoe er later gaat uitzien. Ik geef vaak lezingen, ik zeg vaak, de technologie is zo snel gegaan. Gaande van sociale media tot hoe informatie tot ons komt. Dat er geen morelige invulling aan kan worden gegeven. Dat daarom, die woede halven bij mensen ontstaan, dat ze de wereld niet meer kunnen volgen, die meer weten wat een plek is in die wereld. En nu al, dus laat staan dat het zich verder ontwikkeld moeten we daar niet meer bij stil staan, moeten er niet meer etici zijn, zoals u die zeggen van pauze. Dat denk ik wel, maar dat is wel deel gepeurd. Dus ik kan verwijzen naar bij voorbeeld. Ik denk ook toober, is er een statement on superintelligence? Het schreef een heel kort statement, zonder teken door een 130.000 tall mens een wereldwijd, academische, wetenschappers, politici, bedrijfsmens en mensen het bedrijfswillet aan zijn voort. Mensen die proberen die inzetten op publieke bewustwoording. Rond de huur, maar ook de komstige voor en nadelen van artificee entigenzien. Dus wereldwijd denk ik dat er een beweging aan het ontstaan is, dat er ont publiek bewust en aan het ontstaan is. Dat is heel veel van, want ik denk dat we moeten inzetten, maar ik denk ook dat we moeten voorzichtig centrenatie van de toekomst van artificee entigenzien. Dat voorijs natuurlijk vaak politieke sturing of intervenzien. Die politieke sturing, die is natuurlijk voor een groot stuk, die kan worden gestimuleerd door publiek bewust en dat publieke bewust. Over het risico's zijn kan kanwapens niet verleeren. Dat is natuurlijk een incentive voor politici en beleid. We hebben wel twee keer gebruikt die kanwapens. Die om te bestefen "Holi, holy shit". Dit moeten we aanbanden leggen. Ja, maar er is toch wel een. Mensen zien het z'n raar, we willen het eerst het effect zien. Voor alleen we gaan regioleeren, dat we de politie hebben er vormd. Het is bij de haargetrokkenmiddag voor, maar het is wel zo. De kernbommen zijn gevallen, voor dat we bestefden "Holi shit", wat is dit? Minder dramat is sociale media, daarvan zie we vandaag heel veel voorgen, maar ik raad je wel daarop voor opnieuw vernagedacht. De hele batterij aan gedragspsiegelogen die bij meta aan de knoppen zitten, en die bepaalt er wat kijken, hoe kan we die directs veel mogelijk aan het zijn? En dan komen we aan de economie. Projekst ziet 26 en een keer in Amerika 2500 miljard dollar de EI-industrie. Dan mogen nog 130.000 etici staan wapperen met gelak. De Elon Musk's van deze beeld zijn daar niet mee bezig, denk ik. Het gaat over geld, het gaat over de eerst over de zijn. Klopt. Ja, dat klopt. Maar niet de men, die investeringen daarvan, denk ik ook dat is geen slecht idee. Op de vlood dat de cashflow dat veroorzaakt is, denk ik, geen slecht idee helemaal niet. Alleen, er zijn onvoorzien en onwensijke neven effecten die kunnen plaatsvennen. En daarvan denk ik dat een beleidsmakers geïnformeerd door mensen uit bedrijfsleven, geïnformeerd door mensen uit universitaire wereld, en denk dat die krachten moeten bundelen om druk uit hoeven, zodat we dat zo veel moeilijk kunnen proberen met die geëren. Dat is er voorbij 100 jaar heel vaak gebeurt. Er zijn heel veel dingen geweest waarvan we gezegd hebben kijken of het blik dat die mogelijk was, dat dat heel veel technologisch vooruitgang zou kunnen bieden of dat technisch gezien moeilijk was. Maar dat beslisste hebben we kijken. Op basis van druk van buiten, of hebben we dat niet gedaan? De geloof nog altijd dat de mens dat kan mitigeren, verzachten, milderen, maar hoe meer intelligensie van niet menselijke zorter komt, hoe minder mensen gaan kunnen mitigeren. Want u schrijf ze als aan uw boek dat ei de democratie kan doen, wankelen, holen. Wel, dat denk ik dat het moment heel ook al doet. Een democratieobanklen, dat is wat Europese, wat wij in Europa, wel degelijk al ondervennen, dat heeft er niets mee te ediaan, maar dat zijn taloze bots van uit Rusland die automatiseerd om de juiste informatie verspreiden, om hier onze publieke ideeën, of onze ideeën oplossen schroefen te zetten, om dat informatie te verspreiden, zodat wij gespluiten, raken en zo voort. Dat is geen toekomst, dat heeft in het verleden plaats gewonnen, denk ik aan de Brexit, denk ik aan Trump. Dus dat is alla reden om te zeggen stop de persen even nadenken en even reguleren. Ik denk het wel. Zeker ook als je weet, bedoel je als je kijk naar het democratie, de barometer, er zijn maar een dertigte landweeldwijd, waar we echt in een gezonde democratie leven, dat is de minderheid van de wereldbevolking die een democratie leeft. Als ie daar nu is, en in het verleden al een onwensijke effecting heeft haat op de democratie, dan kun je vermoeten dat dat ook in de toekomst het geval gaat zijn. Oké, goed, dat hebben we nog niet gehad kopen over. Er zijn ook klimatologische gevolgen, die data centers, die zuipen, maar onder duizenden interrogaten voor de goede interesse, denk ik toch. Ja, toch? - Ja, denk ik wel. Ik ben zo met m'n broer, is Stijn, is ook een obetenschap optimist. Die zegt ook van, dat is net het goede aan die artificeerend legendie. Dat gaat dat wel op los. - Ja, dat denk ik eigenlijk ook. De Björnzact is juist, ik heb het in boeken beschreven, er is heel veel elektriciteitswater die nodig is voor het traiende haar van die data center, van de 10.000. Ja, één van die grote data centers, bijvoorbeeld dat staat gelijk met de elektriciteitsverbruik van 100.000 gezinnen. - Ja, dat is allemaal waar. 1000 de centra. - Ja, dat is waar, maar in dat opzicht ben ik pijnna ongebrelde techniek optimist. En er zijn van, ik denk dat we dat aan de hand van onder andere artificeerend legendie gaan aanpakken. En dat is ook al wat plaatsje vervonden in China, vorig jaar deze tijd. Ook precies op de dag trouwens heel interessant op 20 januari in ioguratie. Hij heeft diepsieke Chinese bedrijf, Chinese, mevel diepsieke. En dat blijken heel energieafficient. Er werd wel een helwater nodig. - Ja, maar Westwater gezegd dat er in de kelder een videatiepsant meedraaien waren bij diepsieke, dat werd niet helemaal geloofd, dat het zo energieafficient was. Ja, maar wel in ieder geval financiëel, veel minder investeringen nodig. En dat het niet, dat het niet ideale wereld was dat weet ik ook wel, maar het is niet meer het niveau van GPT-5 en hoeveel het broek has gassen dat uitgestoten wordt om dat van de laatste model te gaan zijn. Ja, en klaas er iets heel interessant, maar zijn jullie een beetje te optimistisch nu? Nee, nee, wacht, was een mooie friggen. - Ja, ik denk nog. Het menselijk brein waarop die neuralen netwerken zijn gebaseerd heeft twee kilo calorie nodig om een heel complexe berekening te maken. Diezelfde berekening heeft op dit moment twee meegabap nodig in een date bij wijze van spreken. Dus als die technologie nog veel meer op het menselijk brein kan lijken, dan wordt het precies wel als het panneetige die tijd niet om zeep is. Ja, ik denk niet dat dat gebeurde, maar ik denk wel dat we AI kunnen gebruiken omdat energieprobleem aan te pakken, maar ook AI om verder onderzoek te doen naar meer elektrische wagen zijn ze voort. En dat is wat vandaag dat ook al gebeurt. Dus die kleine ecologische, van die reale, ecologische kost die geparge met het ontwikkelen van AI, die denken dat we naar beneden gaan halen. En tegelijkertijd denk je dat we die kost moet inkalpleeren aan haar varen omdat precies diezelfde technologie kan gebruikt worden voor meer elektrische wagens bij voorbid. Of kernfusie wordt beter en beter door AI om er tussen. Heel wat van deze gesprekken eindigt met de vraag, hoeveel procent kan's is het dat AI als mensheid gaat uitroeren? En je hebt mensen die 99,999 procent zeggen. En je hebt mensen die zeggen, verdien af 15 procent, wat ik ook al best veel vind. Wat is jouw mening? Ik ga er. Ik ga er weer niet op antwoorden, omdat in het lijkt mij gewoon echt zo wat duimzuigen. Een beetje, we hebben geen data waarop basis waar we ons kunnen stullen om iets te zeggen over de toekomst. Maar als heel veel collega's AI-wetenschappers worden die vragen voorgelegd en zijn heel veel experten die zeggen dat gaat over 2 procent, 3 procent. Wat is al vrij veel? - Dat is veel, hè. Dat is heel veel zelfs. Dus ik denk zelfs al zeg je nog maar 1 procent of 0,5 procent. Dat is al voldoende om daar mensen meer aandacht aan te besteden om daar mensen geld te investeren en onderzoek naar die thematiek. Dat is wat we de voorbij 2030 jaar hebben gedaan rond klimatologisch onderzoek. Maar ook bijvoorbeeld rond de NASA heeft heel veilig investeerd en onderzoek naar hoeken we vermijden als de ruiden inslaat, waarschijnlijk het schraat van die inslaatjes veel kleiner. Toch was dat voldoende. Ik denk dat dat voor een stuk te maken heeft. Maar voor een stuk is dat een commetenschang aan als de ruiden inslaat. Dat kunnen ons visueel goed voorstellen, we begrijpen wat dat is in zeker eens zijn. Maar dat is anders, natuurlijk bij Sweet's, vaak nog abstract zoals artificial intelligence. Dus dat zou een van de factoren kunnen zijn, waarom die zwaar investeringshofen in onderzoek naar die thematiek, hoe groot schallige problemen kunnen vermijden. Dat dat nog onder de radar zit, denk ik. Ik heb nog een vraag. Mijn moeder wordt dit jaar, het is van 39,87 jaar. Voor haar is dit allemaal grote algebra. Hoe moeten veel mensen leven in een wereld die door AI wordt beheerst, waar machines zijn te woord staan als een klacht hebben en ze voert? Ik maak mij daar zorgen over dat mensen te veel mensen zich niet meer thuis zullen voelen in die wereld. En dat dat ook weer hoeven van woede en ontevredenheid gaat veroorzaken. Ik denk eerlijk gezegd dat ook nogmaals ondanks mijn ongebreideld, vaak toch ongebreideld technoptimisme, dat we erop heel vervlakken houden vooruit, maar tegelijkertijd moeten we die zorgen heel reëlen nemen. En ik heb geen zote advies in dat ik kan geven, en ik denk bewustwoording en kennis verhaaren dat dat al een grote stap is, dat we ons blijven informeren. Wat dat mensen zich blijven informeren om het moment heel haan is. En dat er ook mensen zijn die over dat soort van thema's publiek spreken om daar bewustwoording te creëren van wat positief, maar ook negatief aan het gebeuren is. Als je ziet hoeveel eenzaamheid sociale media al hebben te weegebracht, dat is nog maar het begin van alles. We hebben zo behoefte aan zorg gaat die zorgen dan ook AI-zorg zijn, gaat een robotje zeggen van Maria de Krijpamper verwerden. Dat zal wel zetten aan te komen, denk ik. Dat is dat al. - Dat is dat al. Wat een veruitzicht. Bjorn, we gaan eindigen met muziek, heb je menselijke muziek mee of heb je iets aan syno gevraagd. Dat is wel dat muziek genereert. Muziek genereert entrouw is heel populairl. Als je denkt van, we gaan in de toekomst ons niet volop storten op AI genereren de muziekwant, weet ik veel wat. - Is er er? Dat bestaat al zijn miljoenen mensen die die bent. En ze hebben vrij dus de cafés beginnen te spelen dat het niks kost. Vandaag in het warenhuis binnen wandelt, is dat AI genereren de muziek? Ik heb iets van een ouwe kompondisch mee, een mandig leefte. Dus een pak, wij gaan 1864 en 1949, wie is het? Richard Strauss. - Oh. Richard Strauss, een duitser, niet te verwarmen met Johan Strauss van de Ween Zowalsen. Die heeft de mooiste stuk muziek geschreven voor de film. Die beschreven wordt in uw boek ook 2001 Space Odyssey. Maar ook niet AI op dat moment, maar wel een computer. Zijn eigen wildtond, hal 9000, denk ik, en de wilde deur niet meer open doen. Dus vandaar op die film, met dat tema van technologie, die eigenlijk, ja, met ons gaat lopen en de menselijke autonomie gaat afremmen. Wil ik toch die muziek van al zo'n spraag, Zaratustra, latenhoren? Het is niet langer dan minuut 45, maar het is ekonische muziek die bij de tema heel erg past. Lodelaar, ik hou je danken om naar de VRT te komen. Ik kun alles succes met uw nieuwste boek, dat is voor de lente van 27, dus dat is nog wel even. Moet er waarvoor leven wij als AI onze jobs over neemt? Voila, en ik maak ook graag reclame voor dit boek. Eja, wat niemand ons vertelt, ook JetGP10 niet. Bjorn, dank u zeer u, thuis, dank voor het luisteren en we zijn er volgende week weer. [Muziek] [Muziek] [Muziek] [Muziek]