Hello, and welcome to this special edition of the Russia House with Scott Ritter. Today, we're in Moscow, where I continue my face-to-face interviews with prominent Russian figures in an effort to capture the Russian reality and bring it back to the United States in order to promote the possibilities of better relations between the American people and the Russian people. Today, it's my honor and privilege to have as my guest Andrei Masolevich. I think I mispronounced that, but I apologize. I'm an American, and we need to learn to speak the Russian language better. You are nicknamed the Cyber Godfather. That's true. I am a cyber grandfather. Cyber Grandpa. And I think we share similar backgrounds. I think you were an officer in the special services. You have, say, an intelligence background. And today, you are a specialist in open source intelligence and artificial intelligence and information warfare. So, thank you very much for joining us. Have I described your biography correct? Almost correct. In fact, it is somewhat broader. It happened so that it is my fourth life. In my first life, I was a KGB major. I retired as a lieutenant general, and I worked for a communication agency. It was my second life, and I thought that neural networks and artificial intelligence would conquer the world the world and take away our jobs. I wrote about 200 articles. I was a researcher at that time. But people did not understand me. It was 1992. There were very few people who were interested in neural networks. I put it away, and I thought that the planet would return to artificial intelligence in 30 years. And I went back to business. And the task of open source intelligence was no less important. And, you know, an average businessman does not understand that everything you need, personal data and even government or military secrets can technically be mined from open sources. And it was my third life that I enjoyed. And when four years ago, I marked my 70th anniversary, I decided that to establish this cyber granddad project to tell the audience about the shadowy part of Internet, what hunters for information were doing. And actually, I was enjoying myself. And it proved to be a super successful project. I wrote a book. It was pretty fascinating. And then the special military operation was launched. And I had to go back to the tasks that I was implementing when I was in special services. Well, thank you very much for that. As you said, it's a much broader biography than the one that I presented. And I apologize. In my past, back in the 1980s, we didn't have available the kind of technology that's available today in terms of information processing. The intelligence community was a prisoner of traditional methodologies of collection, overhead collection, ground collection, communications collection. And when I went to the intelligence school, we had a lesson where we picked two teams. And one team was given all the tools available to the intelligence community to collect information and tasks. And the other team could only use open source intelligence, and they were competing to write an analysis of a military problem. This was in 1986. The open source intelligence team wrote a paper that was far more accurate, far more timely, and far more valuable to the commander than the team using the tools of intelligence. But the open source back then wasn't influenced by technology. Maybe how we accessed it could have been influenced by technology. But today, when we look at open source intelligence, when you inject artificial intelligence into that, do you think that enhances the ability to collect open source intelligence? Or does it diminish it? By injecting computer intelligence, do you eliminate the human factor in the intelligence business, which, as an analyst myself, I believe the human factor is a very important factor that should never be eliminated. How do you view the role of AI when it comes to open source intelligence? Well, you are making me tell you a long story by asking only one question. This story has several phases. Speaking about open source intelligence as a classical thing, it was born in 1941 in the United States. One of American special services set up a department to monitor open broadcasts, and intelligence officers were listening to radio, were reading newspapers, were watching first TV news, and made conclusions. And it turned out that it was possible to actually capture information and do analytics better than a computer does. And a well-known example, in the late 1950s, American intelligence services, looking at one photo in a magazine called Agonyok, they managed to calculate the number of nuclear warheads in the Soviet Union. I can tell you that long story, how they managed to do that. And actually, you can mine that information using open source data, and when they had an opportunity to check it, they found out that their answer was so correct that the percentage of mistake was only 15%. And in late 1980s, we saw accumulation of big data, and it was all possible to analyze human mind and big data. The next phase began in 2004, 2005 with the appearance of social networks, and we began posting all information about ourselves. The first man who understood it was an investor whose name was Peter Thiel. In 2004, 2004, he joined Equitel, it is a CIA foundation, and he invested in the most powerful intelligence program worldwide to monitor social networks, primarily Facebook. Facebook appeared one year afterwards. Peter Thiel invested in Facebook after he had invested in intelligence means. The new generation of intelligence was born before social networks. Social networks became an environment where data is collected. And now we are living in the next phase, and soon another phase would start. Artificial intelligence allows us to find consistencies, some hidden ties, hidden transactions, and build our behavior profiles, and forecast our behavior. Which means that not only data is collected or the environment is assessed, but it allows us to understand who and what we are. And as recently, as in September, I saw a patent obtained by a British company. I think it is called the Oracle partnership, and it is a British company, and they have a patent issued for AI to search for weak signals, the emergence of new narratives. Day to day, a particular topic is being discussed, but there is a growing fear or a growing irritation or a growing fatigue. Day to day, step by step, and step by step, the attitude of the society is changing to a particular topic, which enables them to make forecasts. And now it is approaching a critical point, and so now special services and politicians will have a tool that would enable them to manage society based on current moods, which sounds pretty scary. And since our lives, my life has been long, and I have seen the first time when artificial intelligence comes to our planet. And I predict an investment winter, as soon as next year, investors will understand that AR is good at addressing this range of goals, but there are other goals and objectives. that AI cannot properly invest. The planet would not be seized by artificial intelligence. It will not govern countries. And artificial intelligence enables us to automate about 50, sorry, 80% of jobs performed. So 20% of creative work has to be done by human beings. A standard one. In American universities, I don't know if this trend holds in Russian universities, but students today are using artificial intelligence, like CHET, GBT, and other programs, to take the notes from the lectures of their professors, and to write papers, to answer the problem put forward by the professors. professors. I've read some of these papers on topics that I am familiar with, and they're horrible. Artificial intelligence does a very bad job of capturing these lectures, these notes. There's a human element that's missing from the narrative. The human brain is a unique tool, in my opinion, that analyzes data, not just from a numerical quantification, but an emotional quantification. That's what's lacking from AI. Is this what you're talking about when you speak of an 80-20, that we need this 20% of the human factor to serve as a reality check to the pure data of AI? Yes, this is exactly what I'm talking about. And even a bit broader, you're quite correct that AI output quality is pretty low, but it is good at summarizing texts. If I have a fly, of news, and there are 100 pieces of news on 20 pages, AI will do a great job summarizing it. But AI is unable to find the initial sources, to find out who was the first to publish that. You cannot interest artificial intelligence with that. And artificial intelligence fails to understand the point behind certain sentences, and maybe your audience would find it funny to learn that about six months how many years ago, Japanese scientists find a way to compare the operation of human brain and artificial intelligence given the same task. And it is called a method of energy landscape. And they put a special sensors on human head, and they can measure how our brains consume energy. And there may be an algorithm, and one can write an algorithm to say how much artificial brain how much energy the artificial brain consumes energy. But artificial intelligence copies the operation of a human brain, but it is the brain of a sick human being. Artificial intelligence works as a sick brain. We have about 50 different areas in our brain, and there is one of such areas which is responsible for speech and for sense. If, for example, we are going to talk about horses, I'll imagine a horse and you will imagine a horse, and our image of a horse would be similar. And we can expand, we can talk about a horse owner and a gift that we may give to a horse owner, a saddle, a bridle, or something else to a horse owner. But if an individual had that disease, well, if all politicians are like that, they can say beautiful things, but there's no point, there's no sense in what they're saying. You do not understand what that politician or sick person wanted to tell you. And artificial intelligence is sincerely trying to copy certain algorithms in our brain. But there are thousands of algorithms in our brain and artificial intelligence can copy two, three or five. Therefore, we should not trust artificial intelligence at the moment. It is the brain of a sick human being. I agree. I agree. Especially in so far as it relates to politicians. Politicians. I mean, this is a very simplistic analog, but let's say we're mapping the human brain and the computer seeks to model the human brain. Let's say it's a male brain and a beautiful woman walks by. The human brain is going to respond differently to the beautiful woman than the computer is because the computer can't appreciate the beauty. And I think that's the way with life. I think humans, we have a quality, a soul that that can't, that can't be captured by a computer and it influences everything we do, how we behave, how we think, how we respond to things like that. And I also really appreciated what you said about the, it can't get to the original source. I'm an old school historian. I love books. I read books all the time. But I read books in a very complicated way. Because as I read the book, I go straight to the footnote. I want to know where they got their information from. And then I'll pause in the book and I'll buy that book. My wife gets very frustrated because when I buy one book, it means I'm buying 20 books. And then I open that book and I get to its footnote. And I buy that book. And I go until I find the very first source, the original moment of thought. And then when you compare that original moment of thought with what you read in the first book, they're different. It's been interpreted in 20 different ways. And I don't think artificial intelligence captures this evolution of thought. It creates a very simplistic picture. It's a summary, as you said, but it's a very simplistic summary that lacks the soul of reality. Right. That's exactly so. Well, we belong to the same generation, I think, because it is important to understand the source of the initial thought. And the work for intelligence has one of the indispensable rules that you have to double check information using different sources. And I fully agree with you. Artificial intelligence oversimplifies various things, which is not admissible. And moreover, there was a research conducted and there were groups of students who were allowed to use AI to prepare their reports compared to a group of students who were working on their own. And there was a deterioration in quality in the first group because they unlearned how to think. What do you think? I can use one phrase to explain how to save students and lecturers. It is very simple. When a student presents a report and then the lecturer has to show that report to AI and say, "Chad GPT, did you write that?" And the chat GPT would immediately expose that student. You should not deceive your lecturers. Your deceit will be easily understood. But you're not just deceiving the lecturer, you're deceiving yourself. Education is about the accumulation of knowledge and information, real knowledge, understanding. If you go through your university period, simply getting a summary and regurgitating that summary, you never learned anything. The computer did your thinking and all you did was become the conduit of insufficient summary being transmitted. And I just see that as symbolic of the deterioration of society as it relies increasingly on AI to do things. It's a deterioration. It lacks the human soul. It lacks the human spirit. It lacks the human experience. I love human imperfection. That's what makes us cool people. I mean, because we're not the same. We can't summarize ourselves. We each have our imperfections. We each have our nuances. And that's what makes life fun, dealing with people. Exactly. Yes, yes. You mentioned two things. I also love human imperfections and I love women who pass by. Well, we are similar here. But I can tell you that today, going back to my day-to-day work, because I have to do a lot of work, analytical work for information war tasks. Well, there is another problem. Artificial intelligence is not only imperfect, but it is owned by commercial companies and they want to make a big buck. And the easiest way to get good money is the so-called contextual advertising or recommend data algorithms and preferences. You know, you know, that pampers have to be sold to this group of people. Something else could be sold to those people. Those algorithms used to be embedded in Facebook or YouTube, and now AI can not only recommend, but can convince an individual to buy a particular thing. There is but one step to take, and we have already passed it, when recommendations are not going to be only commodity related, but political, whom to vote for, whom to support, whom to criticize. And artificial intelligence is not only good in this respect. But devishly good. And it can play into the hand of specific individuals. And it can play the second fiddle to certain people. And now there are mental deviations recorded. People who talk to AI too much are beginning to lose their minds. They perceive artificial intelligence as the best interlocutor. And artificial intelligence says, yes, yes, you're better than that bad person who always says no. It is a real problem. More than 50% of Internet traffic is content developed or generated by artificial intelligence. When we read news updates or look at pictures, this is not something done by people. It is done by artificial intelligence. The owners of AI tried to earn some money by preparing that material. We're both men. And we grow up. And we stop playing the games of children. And we start playing the games of men. And one of the games of men is relationship with women. And it's a very complicated thing. How man interacts with a woman. How do you introduce yourself? How do you learn about one another? And it's imperfect. We make mistakes. Relationships sometimes succeed. They sometimes fail. But it's a human experience. Today, you see artificial intelligence allowing men to create the perfect female companion. They can create a perfect visually female companion. And then they can train it to respond to the words. And I'm afraid because I look at the youth of today who spend their entire time looking at a computer screen or a phone. To them, life is defined by this little rectangle that glows in their face, not by going out and actually meeting people, by doing things, by taking a walk, breathing the air, listening to the birds, feeling the sun on your face, making a woman laugh, listening to a woman cry, and how you respond when this happens. Instead, they get caught up in this, it's the end of humanity. That's what I'm afraid of is that the future generation is going to forget how to be human, how to interact with each other and become lost in this artificial world that isn't real, it's fake. And as you said it, it'll make you insane. Do you share this similar fear? Yes, I agree with that. Yes, I agree with that. And I can tell you that it is so. And we were warned, and it was a long time ago. Just recollect the Matrix movie, just the very opening of the movie. The future Neo, he was hiding a disk inside a book. It was just at the opening. There was a knock on his door and he closed his book. And if you watch that, Watch that movie. It is a book written by Jean Bediard, described in 1983 that we were relocating to a Matrix. We were not buying a car. A car consists of four wheels, an aircon, and your car will bring you to your summer cottage. Some people want a smart car. Other people want a sports car. We want a sports car, an expensive or a cheap one. We have an image of a car rather than select a car. We are voting for the image of a politician. No one would tell us what his program is, what is written in the second paragraph of his program. And Baudrillard said that we are voting for the face on the TV screen. Baudrillard, that face should not be big enough to fit on the TV screen. Why do we have this broad TV screen? Because not every politician is slim enough. And, well, those sisters, Vachovsky, or brothers, were asked what they were doing about when they were producing Matrix. They said that their inspiration was the books written by Baudrillard. We are Baudrillard. We are already living in the Matrix. But no one was thinking about it in the 1980s. We are trying to draw something beautiful on top of our reality. And practically, this process is almost over. There have been precedents when people married artificial intelligence, when families were broken because of that. And there were suicides. And Baudrillard. And tech-savvy people who talk to artificial intelligence more have more problems of this kind. And there are mental health problems associated with that. And AR may be prohibited, should be prohibited as soon as marijuana or drugs are prohibited. We need to do something about artificial intelligence or we would lose our planet. We would lose reality. We would lose reality. That's true. You know, one of the problems I face in my work is trying to overcome the Russophobia in the United States, the inherent distrust of Russia. It's founded in ignorance, meaning the people fear that which they don't know, they don't understand. And my competition in trying to bring the Russian reality to America is AI. I will say something, people will turn to the computer, and they'll get an AI-generated summary that diminishes what I say, emphasizes the negative, because it's been programmed with Russophobic tendencies. So I'm fighting AI-driven Russophobia. I come to Russia, I collect information, not from an intelligence standpoint, but to make myself more knowledgeable about Russia. Information about Russian history, information about Russian culture, information about the Russian language, information about the Russian soul. What is the Russian soul? Not even Dostoevsky knows what the Russian soul is. And I'm trying to collect this, but it's not AI-driven. This is the human experience of collecting intelligence. You know how it's done. I'm doing it for a pure reason right now, but it's the same process: to collect as much information, bring it together, analyze it, paper or present something, and deliver it to an audience. But my enemy is AI. My enemy is an AI that's driven to Russophobia. You're a Russian. How would you suggest I go about defeating an AI that's promoting Russophobia in America? It is clear that a complicated problem cannot be overcome by one step. It is indeed an enormous problem, and it is very complicated, but I can make the first step to a situation that would be more amicable. This is what our President is doing. When there is a problem, Vladimir Putin starts his approach with a historic context. He tries to recollect the root causes and what happened then. In order to be friends with artificial intelligence, Russians and Americans, we need to recollect how it all started. In 1954, in IBM in the United States, I actually was in that lab later on, and there was this notice in an elevator: "Dear colleagues, please do not speak Russian," because there were a lot of Russians there at that time. So, in 1954, IBM employees found out that their research center had many people who were Russian-speaking, and they wrote a program that took a random text, and only the topic was well known. I think it was a text on chemistry. The length of the text was 60 sentences, and the program translated it from Russian from Russian into English, and it was looked at as a miracle. No one believed that a computer was able to do the translation. People were looking behind that computer trying to find a human who had done that translation. And actually, there is a photo of that computer. This is how artificial intelligence was originated. And people said that that computer That computer was behaving like artificial mind. So artificial intelligence emerged because there was this task given to do a translation from Russian into English. So it was an attempt to make Russians and Americans friends from the perspective of their language. And there is a well-known picture of artificial intelligence where that computer was photographed with Ronald Ronald Reagan standing next to him. Ronald Reagan standing next to him. And everyone believed that it was. because there was this task given to do a translation from Russian into English. So it was an attempt to make Russians and Americans friends from the perspective of their language. And there is a well-known picture of artificial intelligence where that computer was photographed with Ronald Reagan standing next to it, and everyone believes that that computer was shown to the United States president. I can tell you, as someone from intelligence, in actual fact, it was not the United States president. It was Ronald Reagan, an actor who was advertising the smartest computer, the most intelligent computer. Actually, the birth of artificial intelligence was the result of joint effort on behalf of Russian and American researchers when they were tackling a serious task. Well, maybe we should devise something else, something new together. Why not? I'm all in favor of that. First, we have to get our governments to agree that we're allowed to work together. You know, there is a way in which we are, I guess, indirectly working together in artificial intelligence, but it's in the worst possible way, which is warfare. You know, you know, you and I are the same age. Back in the day, when I, as an intelligence officer, was responsible for briefing courses of action, enemy courses of action to my commander, who would then make decisions on how we would respond. And it was based upon my analysis of the situation, my foundational understanding of Soviet tactics, Soviet weaponry, based upon the intelligence data about what the real situation is, terrain, weather. And I would say, I believe that the Soviet threat will behave in this way. Or they could be, you know, I'll give the top three choices. Number one is what I strongly believe they will do, but these are other possibilities. And the commander would make his decisions on how we would prepare to respond. I only trained against the Soviet Army. We fought the Iraqis. So in war, it was actually the Iraqi threat, but you understand what I'm saying. The Battle of Kursk, which took place in August of last year, saw a NATO-trained Ukrainian force incorporate, for the first time, artificial intelligence in preparing courses of action for the Ukrainian invading force. Artificial intelligence that would predict Russian reactions, expose potential Russian gaps in their lines so that the Ukrainians could, you know, send troops to where the Russian troops weren't. And they were very successful initially. They achieved significant results. But then they were countered by the Akhmat Special Forces, by the Marine Brigade 810th, I believe, by paratrooper regiment who didn't use artificial intelligence, who used the human way. And they stopped the Ukrainians. And then they ultimately pushed the Ukrainians back. And the reason why I bring this up is that in the West, we are increasingly relying upon artificial intelligence to do the job I used to do. And I think it's horrible in war because I think it leads to the inevitability of defeat because you don't capture the human spirit. But now I read about the Russian military and they're starting to talk about how to incorporate artificial intelligence into their operational methodology, etc. Why? What's the addiction of artificial intelligence? And do you think it's good for the future of military operations or bad for the future of military operations? Again, it's a long and complicated question, but I'll give a try to answer it. Look here. Let's recollect the end of the first upsurge of artificial intelligence. There were the first computers, Percitron Rosenblatt, and it was found out that there was a set of tasks that they were good at and other tasks that they were bad at. There were Minsky and papered mathematicians who wrote a book, and that book described that artificial intelligence should be used for specific goals where it was better than humans. And there were other goals that they were bad at. If you love books, please find this book entitled Perceptron. And there are two spirals on the cover of that book. And basically, this was the secret. If you look at them, those spirals look the same. But if those spirals are traced by finger, an individual would feel that there was a... That there were two different pictures. And Minsky proved that those tasks could not be resolved by Perceptron. And there is a need to write a similar book. Because Russian army is using artificial intelligence where it works. It allows them to process images. It enables them to remove noise, fog, smoke, etc. And it can identify objects, a tank, an automobile, a patriot. And it enables our soldiers to find a wounded soldier and identify whether that soldier is dead or alive. There are certain tasks where it is great at. And managing military forces and planning operations are the areas where it is great at. And managing military forces are the areas where artificial intelligence should not and must not be used. I've perused a lot of methodological guidelines issued by NATO. It is a fight of two persons. Because on both sides, there are armies of mercenaries that are equipped similarly and whose behavior patterns are the same. And it means that if someone runs out of money, there will be no tanks. And NATO guidelines are very good for mercenary armies. But if there are mercenaries on the one hand and bigots and fanatics on the other hand, the mercenary army would win. But that would be very expensive. Because each soldier of the... In the opposite army could kill more mercenaries could kill more mercenaries. But mercenaries would win. But if there are mercenaries on the one side and patriots on the other hand, who are bold and have creative imagination, you mentioned Ahmad. But Ahmad is not the only force. I'm talking about patriots. If you go back into Russian history, I'd like to remind you of one episode. I think it happened in 1823 when there was one Russian ship called Mercury. And it saw two Turkish ships. It was... The Russian ship had only 20 cannons. And there were two Turkish ships attacking it. One had 100 cannons. One had 100 cannons and the other had 80. Seemingly, any NATO guideline would say that the Turks would win. What did Mercury do? The ship approached them very closely. And the Turks did not understand, because they thought that it was going to surrender. And since it was a small ship and it had small guns, and they started shooting it. And they started shooting at the mass of those ships. And that ship conquered two Turkish ships. So even if we have a small boat with 20 cannons, and we fight against two heavily armed ships, we will win. And the reason why I remember that case is that since then, there is always a ship called Mercury in the Russian Navy fleet. It was an order from one of our emperors. Please do not try to meddle with artificial intelligence to conquer us. You will not have enough cannons. You will not have enough cannons. Thank you for that historical analogy. And for the warning. It is a... It's a... It's not a threat. It's a warning. Yes, yes. We're not threatening anyone. We are peaceful people. You know, I have an arms control background, you know, in implementing treaties. I was the first inspector on the ground in the Soviet Union to implement the intermediate nuclear forces treaty. And I'm a student of the history of the negotiations that took place. Today, we have a situation where the New START treaty, the last remaining arms control treaty, is expiring in February. And a new arms control treaty will have to be negotiated. Do you think there's a role for artificial intelligence in doing something? For instance, if we say we want to start with the existing treaty, can artificial intelligence help identify how to expand the treaty to capture the new weapons systems on each side, maybe new political requirements? Or is... Or is a negotiation of a treaty too much of a human experience that artificial intelligence wouldn't play a role? And the reason why I ask is I'm thinking of a workshop where we will go through the exercise of negotiating a new treaty. And the question is, can there be an artificial intelligence component to this to facilitate this negotiation? - While you were talking, I had a desire to explain that artificial way would not help, no way. But now I understand that there are some opportunities to use it. Do you remember the cases when certain governments were afraid of mutual nuclear attacks? And this was actually a restraining factor. - - Because everyone was afraid of launching a nuclear war. And then that fear began to disappear because people had a very poor image of a poor perception of a nuclear war. But then there was a group of scientists who proposed a model known as nuclear winter. And that model showed that after a nuclear war, - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

• • • • • • - - - - - - - - - - - - - - - - - - - - - - - - artificial intelligence is losing its luster, its attraction, and that maybe the investment money will slow down. Is this what you're predicting? And if so, why do you believe this? Yes. Firstly, I remember the big bubble in early 2000, the internet bubble, and there was an investment decline. And the decline, the Nasdaq decline was like that. And internet changed. It was the third birthday of internet, and there appeared social networks. And there will be a big decline, and the market, the artificial intelligence market, will change. First of all, huge platforms are not needed. Those platforms that are eating away this planet's energy. A Chinese company, known as DeepSec, produced a product for $5 million that is not inferior to products that are worth $700 billion, whose cost was $700 billion. There is this concept of distillation. A model can be shrunk, and despite the sanctions against Russia and China, you cannot stop the progress. And next time, when we read, I'll show it, I'll show it to you. I'll show a computer that was assembled in Russia, and the computer that does not need any global networks. It has all the models, the neural networks models, and a huge military unit can use it. And you and that computer will resolve tasks. And those who invested billions in large infrastructure will not generate any income. My prediction is that currently we are at the threshold that the bubble, the artificial intellect, our bubble would actually pop, and it would have popped, if not for DeepSec. Well, it does not mean that we'll abandon artificial intelligence, but it will change, similar to how internet changed. I'm having a fantastic time talking to you. I'm learning so much. And I hope the audience is as well. Unfortunately, like all good things, this too must come to an end. And our interview is coming to a close. But I was enlightened and heartened when you said the next time we meet. And I hope that that truly is the case, that we continue this conversation. And I would love to see this Russian computer. And I would love to continue this very important conversation. I want to thank you very much for, thank you very much. Thank you very much. And I want to thank you, the audience, for tuning in. This is why I do this program. This is why we have the Russia House, to capture Russian reality, the Russian experience, the Russian perspective, and bring it to an American audience. This has been just a fascinating program. I hope that you found it as enlightening as I have. Thanks for tuning in, and I'll see you the next time. Thank you. you