A Harvard-Born Geopolitical AI on Taiwan, Russia, and Why Japan Thinks Differently About Uncertainty
PIVOT 公式チャンネル
Most forecasting asks whether something will happen. Will China invade Taiwan? Will Russia escalate? In this PIVOT interview, Tinatin Japaridze, director of geopolitics at Anadyr Horizon, argues that this is the wrong question. Their company runs simulations of decision-makers' choices and produces many possible futures instead of a single prediction. Japaridze believes this approach fits Japan unusually well, because, in their view, Japan is comfortable preparing for several outcomes at once instead of trying to make the future certain.
The conversation covers four areas: Japaridze's route from music into geopolitics, how the company's engine works, its findings on Taiwan and on Russia and Japan's Northern Territories, and what Japaridze thinks Japan, and the United States, should take from all of this.
From a touring singer to the UN, Columbia, and Eurasia Group
The host opened by admitting confusion about Japaridze's background: press secretary for New York City, UN radio host, Eurasia Group analyst, and professional singer. Japaridze said the host was not the only one confused, and described a career with many pivots.
They started in music very young and were sure they would be a singer. They began touring professionally at 14 in France and later worked in the United Kingdom. Japaridze described themselves as a kind of "export" from the post-Soviet space to the West, at a time when Georgia, Russia, and other countries in the region were opening up. They then moved to New York and worked in musical theater. That work included songwriting for an off-Broadway show called Matryoshka, about the multiple identities people carry, a theme Japaridze connected to growing up with many identities.
Around then they were asked to write an anthem based on the United Nations Charter. That song became their "ticket to the United Nations." Getting a song approved by the UN takes a long time, and during that process Japaridze began writing about the UN for a New York newspaper and eventually became a UN correspondent. For a while they led a dual life: journalist at the UN by day, in the studio at night.
The next pivot came over dinner with friends who studied and taught at Columbia University. Japaridze had dropped out of college at 16 to tour and had never been a student. They applied and got in, studying part-time while still rehearsing and performing. After the first year they chose academia and left the show. Their studies focused on Russia, Russia–China relations, and especially the tech sector, over four to five years. After graduating they worked in New York City government and then spent several years on the Eurasia desk at Eurasia Group. When Russia launched its full-scale invasion of Ukraine, they focused entirely on covering the war, and that was when they decided geopolitics was where their heart was. Music still matters to them, but they said geopolitical risk and technology are now impossible to ignore in daily life.
What Anadyr Horizon does: simulating decisions, not events
Japaridze described the move to Anadyr Horizon as a chance to step back from one region and look at the whole world, with a strong technology component. The company analyzes geopolitical risk through simulations. Human analysts still do desk analysis, and the company adds its own engine, North Star, which Japaridze described as an "agentic systems intelligence powered engine."
The core idea is to change the question. Instead of asking "Will China invade Taiwan?", the engine looks at the signals it collects and maps out pathways that could materialize as the world evolves. It then estimates which of those pathways are more likely than others. It does this by simulating decisions. Japaridze's reasoning is that the geopolitical world turns on decisions made by political leaders and by leaders in the public and private sectors more broadly. The company builds what Japaridze called "digital clones" of these decision-makers and simulates their choices to see which pathway is most realistic for a given question, whether about Taiwan, Russia–Ukraine, or the United States.
The Taiwan simulation: a gray zone, not a conventional war
Asked for a concrete example, Japaridze described a recent simulation on a Taiwan crisis, which the host called the biggest single concern for Japanese viewers. Most coverage Japaridze has seen asks whether there will be a military, kinetic invasion. The company instead framed it as a choice between war, peace, or a gray zone.
According to North Star, a conventional war is less likely than a non-conventional, hybrid, gray-zone situation. Japaridze warned against reading this as good news. In their view, the gray zone is unstable in its own way. Sanctions, cyber operations, and other destabilizing factors could produce a long period of instability, with "lots of different black swans and gray swans."
Japaridze said the engine also surfaced something they had not prompted it to look for: Japan's role. Western analysis of Taiwan usually does not think about Japan right away, or imagines Japan joining crisis management only after events turn hot. The simulation instead gave Japan a key role in deterrence, together with the United States and South Korea. The simulation generated up to 100 possible futures. Japaridze said a strong Japan–South Korea–US alliance appeared in every one of them.
What to do with that: preparing for pathways instead of predicting China
Asked what can be done more specifically, Japaridze said the first step is to stop thinking in a war-versus-peace binary. Their simulations usually produce two or sometimes three broad pathways, each containing clusters of possible futures. Trying to change what China will do is, in Japaridze's words, impossible. The realistic option is to prepare for each possible future. That means reading signals and analyzing decisions at the top of the power structure to see which pathway is emerging, instead of waiting to find out whether China has intervened militarily. Japaridze said Japan is naturally very skilled at this kind of preparation.
Japaridze highlighted one finding in particular: Taiwan's own domestic politics. Analysts usually think about China, Russia, and the United States first and Taiwan's internal decision-making last. The simulation indicated that what happens in Taiwan's legislature will be extremely important and very telling about how a crisis could develop.
On what the private sector can do, Japaridze emphasized resilience over prediction. If a hybrid gray zone is more likely than conventional war, companies as well as governments need to strengthen their defenses in several areas:
- cyber operations
- disinformation
- possible sanctions, both Chinese measures and US sanctions on China, as the situation heats up
The goal, they said, is to stay agile, and that applies to the United States as much as to Japan. Countries can do a lot about their own resilience and very little about what other countries do. Cyber operations, economic instability, and sanctions-related restrictions came up repeatedly throughout the simulation.
Russia: surviving sanctions by becoming dependent on China
Japaridze then turned to Russia, drawing on a separate Japan–Russia simulation about the Northern Territories. Russia has been weakened by Western sanctions since the full-scale invasion of Ukraine. Even so, Japaridze said it has kept some agility, imperfectly, by finding other ways to stay afloat despite new rounds of US and EU sanctions.
The pattern across the company's simulations, they said, is Russia growing more dependent on China. That support can look like a strength, but Japaridze sees it as a weakness, because Russia is losing autonomy. They said this has been developing for a long time, since before the full-scale invasion. Russia restricted its internet and its technology sector to make itself more sovereign and autonomous. Japaridze called the result ironic: very little technological development, and more reliance on China and on countries such as North Korea for technology. They mentioned Iran in the context of drones but said they were speaking specifically about innovation. The simulations show this dependence deepening over time.
The Northern Territories: constant pressure, no thaw
Japaridze said the Northern Territories question was not something most analysts examined until recently. After what Japaridze described as Vladimir Putin's visit to the territories in August, attention increased, and Anadyr Horizon decided to simulate it.
The results resembled the Taiwan case. The simulation showed no signals of a conventional all-out war. It did show hybrid operations and a deepening division, with Russia, China, and their allies strengthening each other on one side and a stronger alliance between the United States and Japan on the other. Japaridze found the US–Japan side promising. Their point was that beyond official statements about each country's importance to the other, they see real support between the two, "actions do speak much louder than words," especially if escalation becomes possible.
Asked what to expect, Japaridze stressed that crises are interconnected. Russia is heavily committed in Ukraine, and they see no sign the war will end soon, even if there is a pause or a brief, weak ceasefire. Because Russia is stretched thin, the simulations show no new military escalation elsewhere. What they do show is constant pressure, with all actors repeatedly holding back from conventional war. A warmer Japan–Russia relationship "does not seem to be on the cards" in the simulations. Japaridze offered an imperfect analogy to the Cold War. Japan, South Korea, and others in the region are aligning with US help and focusing on deterrence over escalation. Russia, meanwhile, is deepening its tech and innovation ties with China.
Track record: the 50-day ultimatum and the Kursk example
The host asked whether the simulations had anticipated past events. Japaridze said the company watches Iran and the wider Middle East closely, mentioning a blockade the previous year. Since then the engine has flagged both broad pathways and specific events based on the signals it detects.
Their main example came from the Russia–Ukraine war. When President Trump gave Russia an ultimatum to negotiate an end to the war, the company simulated the 50-day period that followed. It then compared the engine's predictions with what actually happened over those 50 days. Japaridze reported accuracy of upward of 80% across the events that unfolded, and said many small and large predictions were "absolutely spot-on."
They singled out one result as counterintuitive, saying many analyst colleagues at other firms got it wrong. When Ukraine pushed into Russia's Kursk region, the common view was that Russia would not tolerate it and would prioritize expelling Ukrainian forces before any further offensive in Ukraine. According to Japaridze, the engine predicted the opposite. It expected Russia to keep pressing its offensive inside Ukraine, particularly around Pokrovsk, betting that Ukraine would be focused on Kursk. Japaridze said that is what happened.
Their explanation for why the engine catches such outcomes is that humans think rationally and assume leaders at the political helm do too, which is not always true. A system fed the right data keeps challenging analysts' assumptions about how people behave in different situations. Japaridze stressed the combination of "human and simulant" intelligence instead of relying on AI alone. The most accurate answer, they said, usually lies somewhere in between.
Will AI replace analysts?
Japaridze's answer was no, for now. Speaking about AI in general and not specifically about the company's proprietary system, they said AI still has limits in judgment and empathy and cannot yet replicate a lot of what humans do. That could change over time, but for the foreseeable future both are needed. Japaridze said they find that comforting as a human being.
The same applies to clients. Governments and companies want a person who can explain what the findings mean for them specifically. Japaridze said they do not foresee the human element being eliminated for a long time, while admitting they might be "overly optimistic." The opposite risk is just as real: desk analysis without AI and technology is no longer enough. The bigger danger, they said, is careless data. People assume more data is better, and Japaridze said that is not the case.
What goes into the engine
Japaridze declined to reveal the company's proprietary data, calling it "part of our secret sauce," but described three components.
The first is expert input. The company works with country, regional, and sector experts. The Japan model was built largely with help from Japan experts, including some in Japan. Every political system governs differently, so the US, Japan, and Russia models cannot be treated the same way without losing important nuance.
The second is data hygiene. Data is cross-checked so it is not contaminated or "poisoned." The company does not just scrape the internet without examining what it feeds the engine.
The third is behavioral and psychological modeling. Because the engine simulates decisions and not events, it uses psychological portraits of decision-makers based on how they have behaved historically. That includes how they behave under stress, when sleep-deprived, or when things are going well, since people are influenced by internal and external factors and do not always decide rationally. Japaridze said what makes the engine distinctive is the combination of all three.
How clients receive the results
Clients, both governments and companies, get access to a North Star portal where they can browse all 100 possible futures if they want. The company also provides written and verbal briefs. Every Monday, Japaridze and colleagues hold an online briefing on the latest simulation, and each week covers a different geopolitical hotspot.
Japaridze emphasized that the company also simulates issues that are not yet in the news. The Northern Territories simulation is an example: it was not a top-of-mind issue, but the engine had gathered enough signals to suggest it mattered. The aim, in their words, is to look at "what is not yet in the headlines, but could hit the headlines" and be ahead of the curve. The company also gives one-on-one briefings tailored to each client's interests and vulnerabilities, whether the client is in government, finance, or the commercial sector.
Why Japan: comfort with uncertainty
Asked why they are bringing the technology to Japan now, Japaridze said they have long been fascinated by Japan and worked with Japanese clients at Eurasia Group. They said the world looks different from every vantage point. The Ukraine war, for example, looks different from the United States than from Japan, and being able to move between those lenses is critical.
Their main argument was that Japan is one of the rare countries that adapts to multiple pathways instead of focusing on a single outcome. Japan has complicated neighbors and several hotspots nearby, and Japaridze named China, Russia, and North Korea as examples. In their view, Japan understands it often cannot influence decisions in Pyongyang, Beijing, or Moscow. What it can influence is how it adapts and strengthens itself.
Japaridze said this reflects where forecasting is heading: "I don't think it's about predictions." What matters, they argued, is the window between the first signal that something is coming and the moment it erupts into the headlines. People often forget that this pre-escalation period is the only time when they can change something for themselves. They see Japan as agile and flexible in using that window to build resilience regardless of what happens.
Japaridze contrasted this with the West. "Uncertainty does not frighten Japan," they said, whereas in the United States and Western Europe people are often frightened by uncertainty and want to make it more certain, assuming more information will deliver certainty. Japaridze said it does not. Japan, they observed, is most interested in and comfortable with navigating different futures, and that is what their company offers: options instead of one clear answer, "unless somebody has a crystal ball."
US clients ask for a percentage; Japanese clients ask for pathways
Japaridze said they could speak freely on this because they have spent more than 20 years working in the United States, and the difference shows in the questions clients ask. An American client, in government or business, tends to ask what percentage chance there is that China intervenes militarily in Taiwan. Japaridze's view of that question was blunt: "quite frankly, who knows." Analysts can try to put a number on it, but that is not the most useful output.
Japanese clients, they said, ask which pathways could materialize and which signals would show whether "arc A or arc B" is emerging. Japaridze noted that this is exactly how the company's technology is built.
Advice for Japanese viewers
Asked for a message to viewers, Japaridze offered three points.
The first was about AI. Responsibility matters, but "there is no stopping this car. It's already going." Either AI will use us, or, for now at least, we can use it ethically and responsibly for our own purposes. Pretending it will go away is naive, and no country can afford that view. Japaridze said this is less about embracing AI as such and more about embracing technology and innovation, which they said Japan has historically been good at.
The second was about history. Japaridze, a long-time student of history, said new alliances are forming and the world is changing fast. History is useful but can become an obstacle when people assume a similar situation will produce an identical outcome. Fixating on past parallels risks missing something new and "invisible." Their example was probability: a 3% outcome is still 3%, and sometimes that is the one that happens, while everyone over-prepared for the 97% outcome because historical precedent seemed to point that way. History, they said, "has many answers, but not all of them."
The third was the US–Japan alliance, which Japaridze said matters "now more than ever." In meetings in Tokyo and Kyoto, they saw an openness to doing more together and many shared interests and objectives. Their closing point was that Japan shows strong alliances do not require giving up sovereignty or autonomy. Allies can strengthen each other while staying autonomous. Japaridze said this is something the United States can learn from Japan.
Our company, Anadyr, focuses on geopolitical risk through simulations. We look at the psychological portraits of the various decision-makers based on patterns that we've observed over time. What is not yet in the headlines, but could hit the headlines, and we need to be ahead of the curve.
In the case of the Taiwan crisis simulation, we had up to 100 different possible futures. We were upward of 80% in terms of accuracy of the different events. Japan is one of those rare countries. Instead of just laser focusing on one outcome, you are able to adapt to different pathways.
What does that mean?
Welcome to PIVOT. Tinatin Japaridze is a director of geopolitics at Anadyr Horizon. Thank you so much for coming.
Thank you so much for having me. It's a pleasure.
You have a great background in geopolitics and security, such as your work at Eurasia Group. And you also have a lot of experience in the public sector and media, such as a press secretary for New York City and UN radio host. And also, you are a professional singer as well. I was confused. Who are you?
Could you introduce yourself again, and especially what motivated you to enter the field of geopolitics?
Thank you very much, once again, for having me, and I promise you you are not the only one who is slightly confused by my background. Sometimes I also have to navigate the many, many identities. As you pointed out, and quite rightly so, I have had many, many interests over the years, and I've taken many pivots in my career. But I feel very fortunate that I was able to build on the different experiences that have helped me to get to where I am and where I hope to get over the years to come.
In terms of how I got into geopolitics, one would say that it's actually a little bit of an accident, because I was in music. As you and I were talking before this interview, I started extremely young. And I was always fully convinced that I would be a singer. It was just one of those things that you know deep down. This is who I am and this is who I'm going to be. But as I started touring professionally when I was 14 years old, I started my career in France.
14.
Later got to the United Kingdom. In a way, one could say I was an export from the post-Soviet space to the West as Georgia, Russia, and other countries in the region were opening up. And then I ultimately ended up going to the United States, specifically New York, where I was able to not just make music, but also start a career in musical theater.
I was doing songwriting and working on an off-Broadway show that was called Matryoshka, which is a nesting doll. And that was very much about the different identities that we embody, and those of us, in my case, who grew up with many identities. And around that time, I also was asked to write a song, an anthem that would be based on the United Nations Charter.
Really?
And that song became my ticket to the United Nations. And I became very interested in the organization. While I was performing there and taking lots of meetings, because you can imagine getting a song approved through the UN is a long process, I actually started writing for a newspaper in New York about the UN. And over time I became a UN correspondent. During the day I was at the United Nations working as a journalist, and in the evenings I was in the studio.
Really impressive.
I had a little bit of a dual life that I was leading. I was moonlighting. And it was really fascinating. I was in my early 20s. I had just arrived in New York. It was very exciting. And for a long time I ended up actually having both careers until once again I made another very unexpected pivot when I decided to go back to school.
What was that?
I was having dinner with friends of mine who were actually both studying and teaching at Columbia University in New York. And they were telling me over dinner about their experiences at school and I got very inspired. And I thought, I've had this wonderful life that I've been lucky to have in the music world and traveling and singing and writing. And I thought, this is all great, but I never got the experience as a student because I dropped out of college when I was 16 to be able to go on tour. And I thought, well, what if I just gave it a try?
Really?
And I did and I got into Columbia University. And at first I was part-time. Again, I had two lives. During the day, I had school and in the evenings, I would go to my rehearsals and have my performances.
So busy.
It was very busy. But after the first year, I had to make a decision. And I thought, well, now I need to choose what am I going to do for hopefully the rest of my life. And I ended up leaving the show and pursuing my new calling, which was school, academia. That got me interested in Russia, Russia-China relations. The tech sector in particular was always compelling to me. And over about four to five years, I fully focused on my studies.
And by the time I graduated, it was of course a very natural place for me to end up. First, I went into New York City local government, as you pointed out. And then I got a fantastic opportunity to work at Eurasia Group, where I spent quite a few years on the Eurasia desk, actually. And around the time when Russia started a full-scale invasion of Ukraine, that's when I was laser-focused on covering the war. Of course, at that point, I decided that this was really where my heart was.
Music will always have a very special place in my heart, but geopolitical risk, technology, it's really impossible nowadays to ignore it and it's such an important part of our day-to-day lives that for me, I feel very lucky that I am able to both work in this field, but also have a hopefully slightly clearer understanding of the different geopolitical shifts that we are seeing in the world. It's both my day-to-day life and my professional life now.
That's a really powerful story. Thank you for sharing the story. You started your career from the music world, then you're finally focusing on geopolitics.
Yes.
As the world changes. Now you're working with Anadyr Horizon, right? What kind of company is it?
Anadyr Horizon has been a new chapter in my life and one that I am extremely excited about because here instead of just focusing on the Eurasian region, I am able to essentially zoom out and look at the world, the macro picture. And that's been a fantastic opportunity, particularly because there is a very strong tech component, because our company Anadyr focuses on geopolitical risk through simulations.
Instead of just analyzing geopolitical shifts and events, we do so not just through desk analysis with our analysts, but also using our engine, our agentic systems intelligence-powered engine, North Star, which helps us to look at the bigger picture and also to zoom in on the different possible futures that we can anticipate instead of just looking at one answer. For example, will China invade Taiwan is a question that most analysts, traditionally speaking, will ask. In our case, using our engine North Star, we are able to shift the question a little bit and instead think about the ways that we need to prepare in terms of the different pathways that could materialize depending on how the world continues to evolve over a period of time.
We look at the different signals that we are collecting that then tell us in turn what are some of the pathways that we can anticipate, which ones are most likely versus others, and we do all of that through simulating decisions. Because at the end of the day, the world, the geopolitical world especially, turns through the decisions that are made not just by leaders, political leaders, but in general various leaders, whether it's in the public sector, in the private sector, and that is exactly what we look at. We simulate through our digital clones of the various decision-makers different decisions that in turn can tell us which is the more realistic, more likely pathway that could materialize over time vis-à-vis a specific question, whether it's Taiwan or Russia-Ukraine or the United States, etc.
This is really impressive technology, but it's still a very, very new concept for most of us. Could you give us some more examples of how it works?
For example, we recently conducted a simulation on the Taiwan crisis.
Taiwan crisis?
Yes.
That's the biggest single concern for Japanese viewers.
Absolutely. And while the question that I'm encountering through different analytical articles, different media coverage about what could happen, it's centered on whether or not we will see a military kinetic invasion. The way that we analyzed the possibility and the probability of some sort of an intervention from the Chinese side using our engine was whether we should think about this as a war versus peace situation or a gray zone. And what we found through our simulation was that actually, instead of anticipating a conventional war, what we can anticipate that is more likely to materialize, according to North Star, is going to be a non-conventional, hybrid gray zone.
What was that?
Now, at first glance, one could think that, oh well, that's actually really good news because that means that there's a lesser likelihood of a full-scale war. Maybe a gray zone is not so bad after all. But the issue is that this gray zone is equally, in a different way of course, unstable. Because there are economic factors, sanctions for example, cyber operations, and a lot of other destabilizing factors that could actually turn this into a prolonged state of a very unstable gray zone area with lots of different black swans and gray swans that could keep this situation very unstable for a longer period of time.
But what makes it also interesting is when, in the West at least, when we look at the Taiwan crisis, we don't immediately think about Japan's role in it. But our engine, North Star, actually, through the simulation, without us even prompting it to do so, uncovered Japan's role in this. Instead of getting involved in the crisis management post factum, if things do get very hot, afterwards, instead, we see Japan having a very important, very key role in the deterrence process.
Deterrence process.
Yes, with its allies, especially the United States and South Korea. We see this very strong alliance that continues over time, and it comes up in every single possible future. For example, in the case of the Taiwan crisis simulation, we had up to 100 different possible futures uncovered.
So many.
Oh my goodness.
So many.
But one of the common threads through all of them was a very strong Japan, South Korea, US alliance that we saw.
What can we do more specifically?
Well, what we can do more specifically, I think, is to actually stop just thinking about the binary, war versus peace. Instead of just thinking about whether or not China will intervene militarily into Taiwan, what we need to think about instead is what are the different ways that the decision-makers can prepare for different pathways that could materialize. Generally, we tend to see two, sometimes three different pathways that could materialize. And within them, we see those clusters of different possible futures that I mentioned.
And I think one of the most important things to bear in mind, and something that I do believe Japan is naturally extremely skilled at, instead of trying to change the future and trying to change or influence what China will do, which is impossible, we can instead prepare ourselves for the different possible futures. And also prepare ourselves through, instead of waiting post factum to whether or not China has intervened militarily, to look at the various signals and to analyze the different decisions at the top of the vertical of power that could tell us which pathway is materializing.
And in this case, instead of just looking at China or even Russia, for example, and the United States, of course, also a very important actor here, we need to also look very closely at the domestic situation and the domestic decision-making inside Taiwan. Because usually that's the last thing that analysts think about. But in this case, our simulation showed that what happens in terms of the legislature inside Taiwan is going to be extremely important and very telling of how the crisis or potential crisis could actually evolve.
You talked about the importance of alliance, but as for the private sector or economic sector, are there any ways, measures to decrease the impact of the crisis?
Well, given the fact that our engine suggests that there's a lesser likelihood of a kinetic conventional war, but rather a sort of unstable gray zone that I mentioned that has a hybrid component to it, resilience is very important. Instead of trying to predict the next move, building up and strengthening the private sector's resilience, not just the governments, but also the private sectors, vis-à-vis cyber operations, disinformation, of course, is very important here. Potential sanctions, both from the Chinese side and the US side on China that could be possible as the situation becomes hotter and hotter, is very important so that we maintain agility.
And by we, I say not just in this case your country, Japan, but also the United States. Because there is so much that we can do in terms of our own resilience and navigating uncertainty ourselves, but very little that we are able to affect in terms of what the other countries will do. There we don't have a lot of control. But I think cyber ops in particular kept popping up throughout the simulation that we noticed, and also economic instability and various restrictions in terms of sanctions, for example.
I will give you one example that was very interesting to me. Russia in this case, and we also looked at a Japan-Russia simulation separately vis-à-vis Japan's Northern Territories, normally we think about how will Russia be able to handle the various restrictions in terms of Western sanctions. And as you have seen, after the full-scale invasion in particular of Ukraine, we saw so many sanctions on Russia, but at the same time, we are observing that Russia is able to somehow, not perfectly, but they are somehow able to maintain some sort of agility in terms of looking for other ways that they can stay afloat, irrespective of the latest wave of sanctions that may be coming its way from the West, both from the United States and from the European Union as well.
Yes, on the one hand, it's been weakened by Western sanctions, but what we are seeing through our simulations across the board is that instead, we are witnessing Russia strengthening its dependence, alliance, but also dependence alliance, in a way, on China. And that could be seen not just as a strength, but also as a weakness, because it is losing its autonomy, but in fact, what's happening is it is becoming even more dependent rather on China.
And that is something that has been a long time coming, because through closing or at least restricting its internet and its technological advancements over the past years, not just right as the full-scale invasion of Ukraine started, but prior to that as well, Russia has been weakening itself. It has been less and less autonomous, even though that is quite ironic because the whole point of restricting the Russian internet space was to make it feel more sovereign, more autonomous. But in fact, what we've seen was that there was very little technological development over the past years. And therefore, it is more and more dependent on and reliant on China, as well as other countries in terms of tech, such as, for example, North Korea. Obviously, there's Iran when we look at drones, but I'm talking purely just in terms of innovations.
And that has been a real problem for Russia. And that is something that we are seeing in our simulations, where
That situation will become only deeper. It will only deepen over time.
That's really interesting. So, could you talk more about Russia-Japan relations, about the northern part of Japan, as you mentioned?
That's very interesting because even recently, it was not something that most analysts looked at. But after Vladimir Putin's visit to the northern territories in August, we started paying more attention to this. And that is one of the reasons why our company, Anadyr Horizon, decided to conduct a simulation on this question and what is likely to happen.
And similar to the Taiwan crisis, it's not so much that we are seeing any signals of a conventional all-out war, but instead, hybrid operations and this deeper division. We're seeing Russia, China, and other allies of theirs on the one side strengthening each other, supporting each other. And on the other side, we are seeing a stronger alliance, especially between the United States and Japan.
So, that is very promising because it's not just that we're seeing statements coming out of the Japanese government and the United States government in terms of how the other country is important to us, but actually what's very interesting is that beyond just words, actions do speak much louder than words. And we are seeing real alliance and real support vis-à-vis one another, in this case between the US and Japan, especially in the event of a potential escalation.
So, given that point, what should we expect to see?
In many ways, given the fact that all the geopolitical crises are interconnected, because Russia is so heavily involved in Ukraine, and at the moment, we are not really seeing any signs that the war in Ukraine is ending anytime soon, even if we see a pause, even if we see a brief, weak ceasefire. That war, unfortunately, does not seem to be ending anytime soon. And therefore, the Russian side is stretched thin. So, in terms of another military escalation elsewhere, we are not seeing that, fortunately.
But what we are seeing is constant pressure, but with all of the actors constantly stopping themselves from engaging in a conventional war. So, yes, on the one hand, it's promising that we're not seeing a military kinetic escalation, but on the other hand, that gray zone, similar to the Taiwan crisis, is ever present. So, in terms of seeing the relationship get any warmer anytime soon, unfortunately, that does not seem to be on the cards, and we're certainly not seeing that in our simulations.
But what we are observing instead is stronger alliances on the two sides. One can think about it in a different way, of course, and it's not the perfect analogy, but the Cold War model of having the different powers come together and strengthen one another. And we are seeing something similar here with Japan, South Korea, and others in the region that are coming together with the help of the United States and very much staying together on the same page to ensure that they think about deterrence rather than any sort of escalation. But then on the Russian side, strengthening of ties, particularly when it comes to tech and innovations with China.
I see. So, we have been talking about future prediction, but I also would like to ask about past events, like how your simulation predicted major events before they actually happened.
Yes. Iran and the Middle East more broadly is one of the regions that we are monitoring very, very closely, as you can imagine. So, even last year, we did see a blockade. And since then, we've been monitoring the space very closely. So, we are constantly seeing not just broader strokes in terms of pathways that could emerge, but specific events that could materialize based on the signals that our engine is spotting.
Same for the Russia-Ukraine war. You will probably recall last summer when President Trump gave Russia an ultimatum to come to the table with Ukraine to end the war. We conducted a simulation on that 50-day period to see and to compare how our engine's predictions would actually align with the reality that unfolded over the 50 days that came thereafter. And it was fascinating how many different small and big predictions were absolutely spot-on. We were upward of 80% in terms of accuracy of the different events that unfolded.
One example that was very striking, that was very counterintuitive because many of my analyst colleagues in other companies actually did not get that one right, was, as you recall, when Ukraine went into Russia's Kursk region on Russian territory. Everyone thought that there was no way that the Russian side would tolerate this, that Vladimir Putin and the Russian Ministry of Defense would prioritize pushing Ukraine out before they thought further about any further offensives on Ukrainian territory.
However, our engine actually predicted something that would unfold, and that was specifically that instead of pushing the Ukrainians out of Kursk, Russia would instead continue pushing its offensive inside Ukraine, specifically Pokrovsk, and not abandon that campaign because they expected the Ukrainians to be fully focused on Kursk. And that was exactly what happened. And there are a lot of different findings that we see through our simulation that are oftentimes counterintuitive. And one of the reasons is that we as humans think very rationally. And we also think that decision-makers at the political helm of power are also rational beings, but that is not always true.
I see.
And that's where I think having our engine, North Star, and in general the right kind of technology with the right sort of data that is fed into it is very helpful because it constantly challenges us as humans in how we think about how the events may unfold and how individuals behave in different situations. So, having both the human intelligence as well as the simulation intelligence, so human and simulant, that combination right now is extremely important without us relying solely on AI and having the best of both worlds, if you'd like, to work together, because there's always something in the middle that ends up being the most accurate, the most telling.
So, you said the combination. So, you do not think that AI will replace human analysts?
For the time being at least, given some of the restrictions that AI broadly—and I'm not talking specifically about our own agentic system, our own proprietary system—but in general, there are of course still restrictions in terms of judgment, in terms of empathy. There's a lot that AI cannot replicate as of yet, at least. But over time that could change. But for the foreseeable future, I do think, and it's comforting for me as a human being, that we still need both. And we need that human touch.
And also, on the other side, when we think about those who are on the receiving end of this information, for example, different governments, different companies whom we service as clients, they also need both because they do need a human being to be able to talk them through the different implications for them specifically. I don't think that we're there to be able to completely eliminate the human factor. And for a very long time, and perhaps I'm being overly optimistic, but for a long time I don't foresee that happening. On the other hand, it is no longer enough just to have desk analysis and not to turn to AI and technology broadly, either.
The risk is and the danger is when we are not careful with the data that we are feeding our machine. Because that's what we oftentimes miss. We think that more information, more data is better, but it's actually not.
What kind of data are you feeding to the AI?
So we have our own proprietary data, of course, which is part of our secret sauce. So without revealing too much of that, we also work very closely with country and regional and sector experts as well. So for example, our Japan model was built in many ways through the help and the support that we got from fantastic Japan experts, including here in the country. So that's very important because it's a very different way of governing. Every system is very different, so we cannot treat the US model in our engine the same way as we treat the Japan model, for example, or the Russian model. So there are a lot of nuances that cannot get lost. So there, of course, we do turn to experts.
We have a lot of data that we also cross-check so that it's not contaminated, it's not poisoned in any way. So we're not just collecting, scraping data off of the internet without actually looking closely at what it is that we are giving to our engine.
And lastly, but certainly not least, we look at the behavioral aspects because there's a very strong psychological component. Since we simulate decisions and not events, we look at the psychological portraits of the various decision-makers and how it is, based on patterns that we've observed over time historically, how they've behaved in different situations, whether when they were under stress, for example, sleep-deprived, when things are going really well, because at the end of the day, humans are humans, and they can be biased by certain external and internal factors. So, they're not just making decisions rationally at all times. So, there are various components, various elements that feed into our engine, but I would say that what makes it very special, very unique is the combination of all of the above.
Your clients are companies and the government, right?
That's right.
How do you convey those simulation results to them?
So, we have a portal of our engine North Star.
Mhm.
And they have access to our portal where they can look, if they want to look at all 100 possible futures, absolutely no problem at all. But also, we have various briefs that we give to them both in writing as well as verbally. Every Monday, with my colleagues, I put together a briefing that is done online of the latest simulation that we've done. So, every week we do a different simulation based on a different geopolitical hotspot.
But we also look at events that have not quite really unfolded yet, and we are collecting signals that it may become important. So, for example, when we looked at Japan's northern territories and the possibility of a Russian escalation there, that may not have been a top-of-mind issue, but we were collecting enough signals through our engine that told us that we need to look at this. It's important. So, we look at both what's in the headlines, but even more importantly, what is not yet in the headlines, but could hit the headlines, and we need to be ahead of the curve.
We do a lot of personal one-on-one briefings for our clients as well based on their specific needs and areas of interest or potential vulnerability, so we can tell them specifically what the implications are for them. So, we look at the big picture, but then we also zoom in. We do tailor our work and our research and our findings based on what it is that the client needs, whether they're in the government, they're in the financial sector, or the commercial sector.
So, now you are bringing this technology to Japan, right? So, why Japan and why now?
I think Japan is a very interesting case study. For me personally, it's always been of great interest. I have always been fascinated by Japan.
Really?
And over the years, I worked with Japanese clients, including when I was at Eurasia Group. So, that to me was always very, very interesting because the world looks very different from every part of the world. So, the war in Ukraine, for example, looks very different from the United States, but it also looks very different from Japan.
I see. Exactly.
To me, that lens is critical. And being able to navigate the different lenses. But in terms of our specific technology at Anadyr Horizon, I think we are really well suited, if I may say so myself, our technology is, for Japan and Japan's needs, because Japan is one of those rare countries where you are able to, instead of just laser focusing on one outcome and the possibility of that one outcome manifesting into reality, you are able to adapt to different pathways.
What does that mean?
For example, if we look at geography and the geography of uncertainty, Japan is constantly having to navigate different geopolitical hotspots and very complicated neighbors. So, we're talking about China, we are talking about Russia, we are talking about North Korea, just a handful of examples. So, there's a lot that Japan has to process, but while it may not be able to, and it fully understands it is not able to oftentimes influence the decision-making that happens inside North Korea, for example, or China, or Russia, it can influence how it adapts to this uncertainty and how it strengthens itself because it's about resilience.
I don't think it's about predictions. I think the future of forecasting is going to be about using this interim time between when we first collect a signal that something is about to happen versus after it's erupted and it's hit all the headlines. There is that period of time before the escalation that oftentimes we forget that it is actually the only time when we can change something for ourselves. And I do believe that Japan is extremely agile, very flexible in using that period, that critical moment in time when it can actually do what it can to be able to prepare itself and to increase its resilience irrespective of what may happen.
So, uncertainty does not frighten Japan, whereas I notice in the West, especially in the United States and Western Europe, we're oftentimes really frightened in situations of uncertainty. We want to make it more certain. We think that if we collect more information, somehow we will be able to attain more certainty, but that is not the case. In Japan's case, what I'm seeing is that it's precisely in navigating the different futures and trying to navigate the different patterns and pathways that Japan is most interested in and most comfortable with. And that is precisely what our company does. It offers the different options as opposed to one clear answer. And that I think is far more important at the moment than for us to be looking for that one answer, unless somebody has a crystal ball, of course.
Is Japan's way completely different from other countries like the US?
It's very different from the US, and I can speak about this freely because I have spent the majority of my life, over 20 years, working in the United States. And I can tell you that that is very, very different. Even the way that I interact with our clients, I get very different questions from US clients versus Japanese clients.
For example, like I mentioned the Taiwan crisis, an American client, whether they're in the private sector, they're in the government, regardless, is likely to pose the question that goes along the lines of, "What is the likelihood, what percentage can you put on the likelihood of China intervening militarily into Taiwan?" Whereas in Japan, instead of asking us that question, which, quite frankly, who knows? We can try to be analytical and we can try to attempt putting some percentages up on the probability, but at the end of the day, what we can offer and what Japanese clients seem to be most interested in is, "What are the different pathways that we could see materialize?" And then based on those pathways, what are the different signals that would tell us which of the pathways, let's say arc A or arc B, is the most likely one to materialize? And that is precisely how our technology is built, actually.
This is very, very interesting. So, do you have any messages or advice to Japanese viewers, what we should think about?
First of all, when it comes to AI and the uncertainties surrounding technology, but specifically where AI is headed and whether or not we should be nervous, I do think that, of course, we need to be extremely responsible, but at the same time, there is no stopping this car. It's already going. It's already started. Either AI will use us or we can still, for the time being at least, continue to use it for our own purposes. We do so ethically. We do so
Responsibly. But pretending that it will go away anytime soon, I think that is very naive, and I don't think any country at the moment can afford to think that way. So I'm not saying necessarily let's embrace AI, but in a way we need to embrace technology and innovation, and that is something that Japan has been fantastic at over the years in terms of advancing and embracing technological innovations.
In terms of geopolitics, now is the time when we are going to see, I believe, new alliances forming. The world is changing very fast. Yes, history is very important. I was a student of history for many, many years. That is something I'm very passionate about, but at the same time, while it is very useful, it can be a little bit of an impediment because we over-rely sometimes on historical patterns, thinking that if something happened already in this particular way, then the next time a similar situation unfolds, that we are going to see an identical outcome. And that is not always the case.
Yes, history is useful. It's helpful, but it is not always the only answer because the fear is, the risk is, that we are going to potentially miss something invisible that is new because we are so laser-focused on trying to make parallels with the past, because the context is of course always changing. So knowing when to look to history and when to think about, well, actually, what could happen that is different? What are some of the black swans, anomalies that we should be prepared for?
Something may be a 3% probability outcome, but it's 3% still. And stranger things have happened where that 3% outcome was actually the one that we ended up with and not the 97% one that everybody comfortably anticipated and overprepared for while ignoring this, based oftentimes on historical precedent, because we think that history has all the answers. It has many answers, but not all of them.
And the third thing I would say is the importance of the US-Japan alliance. I think now more than ever, and we are certainly seeing that on both sides, and I've been very happy to see in my meetings with Japanese colleagues here in Tokyo as well as in Kyoto that there is this openness now. There is this openness to be able to do more together because at the end of the day we are much stronger together than apart. And we have so many mutual interests, so many mutual objectives that now is the time for the different blocs to come together.
And yes, we can absolutely be autonomous without us giving up alliances, and that is one of the great strengths of modern-day Japan that I'm observing that I think is something that we in the United States can learn from. That actually having strong alliances does not mean that we have to give up our sovereignty and autonomy, not at all, quite the contrary. We can strengthen each other.
Very, very important point. Thank you so much for your takes.
Thank you very much for having me.
People.
Yeah.
[music]
Yeah. Yeah.
Article published · Updated