column
The Future of Life+ | Dialogue between Yujin Kitagawa (Yuzu) and Hiroshi Ishiguro
To depict the continuation of the future of life
"The Future of Life," which served as The signature pavilion for the 2025 Expo 2025 Osaka, Kansai, is the basis for the currently running "The Future of Life +" exhibition. Many celebrities were among the visitors, and Yujin Kitagawa of Yuzu, who is participating in this interview, is one of them. In the tour for Yuzu's album "Shinon," which began in May 2026, they undertook the first-ever attempt to tour arenas nationwide with only acoustic performances, attracting approximately 220,000 people. During the performance of the song "Keshigomu" (Eraser) on that tour, Kitagawa actually "collaborated" with an android that appeared in "The Future of Life."
The collaborative project for "Eraser" took a more concrete turn when Kitagawa visited Ishiguro in March of this year and the two talked. The timing was perfect, as Ishiguro and Kitagawa were both in the midst of preparing for "Inochi no Mirai +" (currently being held at Mon Takanawa) and Yuzu's arena tour, respectively. This led to a raw and heartfelt exchange of words between two people who create things, transcending their respective positions.
I want to record this dialogue between two people who are exploring the question of "what it means to be human"—one through robots and AI, the other through the creative process of music.
Interview and text by Masahiro Tanioka
How to understand humans through robots
Kitagawa
It's an honor to meet you. I'd only ever met you as the android professor at Expo 2025 Osaka, Kansai pavilion (laughs). The expo itself was wonderful, but I was particularly moved by Professor Ishiguro's The signature pavilion, "The Future of Life." As I was working on the album and preparing for the tour, I thought that this might be a way for us to collaborate, so I proposed a collaboration between me and an android on the song "Eraser."
Ishiguro
Thank you. Regarding Expo 2025 Osaka, Kansai pavilions, we spent a considerable amount of time and effort creating them thoroughly, so it's wonderful to receive feedback like that, saying that it impressed people.
Kitagawa
How long did you spend preparing for this?
Ishiguro
It took four and a half years. We remade each and every scene many times.
Kitagawa
I see. Since we have this opportunity, would it be alright if I asked about its roots, or rather, its origins?
Ishiguro
here you are.
Kitagawa
What initially sparked your interest in robots and androids, Professor?
Ishiguro
It could have been anything. When I was in fifth grade, my teacher told me to "consider other people's feelings," but I didn't know how to understand what it meant to consider other people's feelings. After all, I'd never seen what other people's feelings were like. They said to think about it, but nobody taught me how to do it. But what I thought at the time was that I would be amazed that I would understand such things when I grew up. I thought, "Adults are so incredible!" However, as I got older, through middle school and high school, I started to realize that nobody was thinking about those things. So I thought, "I'm not going to become an adult!" (laughs), and instead of choosing to become a working adult, I stayed in university. I drew pictures in both high school and university, but it was tough to make a living as an artist, so around that time the internet came to Japan, and that's when I went into the computer world.
Kitagawa
Oh, I see.
Ishiguro
Yes. And computers are structurally similar to the human brain. That's where my interest in AI began, and I realized that there's no point in creating something with just a brain using AI, so I thought about robots, and then I wondered what the ultimate robot would be, and that would be a human, so I decided to create a human. So it just naturally progressed this way. Thinking about "What is a human?" is the same question I had when I was told in fifth grade to "consider other people's feelings," and ultimately, the questions we have as children are connected to fundamental aspects of humanity that even science hasn't been able to answer. When I said at the beginning that anything would have been fine, I meant that I've never understood "What is a human?" since I was a child, and if I could understand it through pictures, that would have been fine too, but by chance I went from the internet to AI to robots, and now I'm trying to understand humanity through robots.
Kitagawa
In that process, were there any moments when you felt you truly understood someone's feelings?
Ishiguro
For example, what does it mean to have a conversation? I came up with this idea fairly early on, more than 20 years ago, and I created a robot that could mimic people. It only mimicked gestures, using motion capture. At the time, it was incredibly expensive, costing nearly 100 million yen, to create a robot that could perfectly track human movements. But it was like a mirror. It could perfectly mimic. However, if the tracking movements were random—by random, I mean the robot would just repeat an action that was done a little while ago—it wouldn't be communication at all. But if you make it mimic about 80% and reproduce an action that was done a little while ago for the remaining 20%, it creates a tremendous sense of communication.
Kitagawa
Does "it feels random" and "recreating actions from a little while ago" mean recreating human movements with a delay?
Ishiguro
No, I mostly imitate them, but sometimes I assert my own style.
Kitagawa
I see. It certainly does seem like the ego is manifesting itself there.
Ishiguro
Communication is made up of 80% empathy and 20% self-assertion.
Kitagawa
Did you know that rule from the beginning?
Ishiguro
This is something we've come to understand through conducting this experiment.
Kitagawa
That's interesting. So that leads to the next step, right?
Ishiguro
That's right. When it comes to the actions and interactions of robots that interact with people, we can improve their performance by considering how to control them in a way that draws humans in. Making eye contact is one example; it would be creepy if they made eye contact 100% of the time, wouldn't it?
Kitagawa
surely.
Ishiguro
Humans don't make direct eye contact. They seem to be, but they're not. There's a balance there, like 80% and 20%, or 70% and 30%. I'm researching how humans work by finding reproducible patterns like that and creating copies of humans.
Kitagawa
What was the state of AI like at the time of the research you mentioned earlier about robots that mimicked human movements over 20 years ago?
Ishiguro
35 years ago, Japan was number one in the world for what was called the "fifth generation computer." Even back then, multi-layered neural networks like deep learning existed. However, they couldn't compute quickly. There's a research institute called ATR, which I'm involved with, and it's kind of like a long-established AI company. At the time, they didn't have algorithms for fast computation, but they were creating some pretty complex neural networks.
Kitagawa
As a layperson, I would think that deep learning is more difficult than simply calculating quickly, but apparently that's not the case.
Ishiguro
That's not it. Originally, neural networks were designed to have an exponentially increasing computational cost, so no matter how fast the computer became, it was considered impossible to compute them within a practical timeframe. But then, an algorithm was discovered that drastically reduced the computational cost, so it became clear that with this, you could just pour money into it and make it bigger and smarter.
Kitagawa
The concept of exponential growth is like the expansion of the universe, isn't it?
Ishiguro
That's right. We give up on things where the computational complexity keeps expanding like this, but we can do it, so we just keep collecting data and making the computer bigger and bigger. On the other hand, when computers were small, all sorts of research was born from them. Because all sorts of ingenuity was required. But now it feels like we only need Transformers (deep learning models that form the basis of current AI technology), and it's all running on almost the same architecture. The only thing left is how much money we spend and how big the computer we use.
Kitagawa
So, earlier we discussed how Japan was number one in the world during the era of the "fifth generation computer," but what's the situation now?
Ishiguro
It depends on the field. Japan has always been and still is the undisputed leader in industrial robotics. The automotive industry is also overwhelmingly dominant. However, in the field of AI, it's difficult to compete. You can't catch up unless you repeatedly invest tens of billions of yen.
Kitagawa
It will be tough unless the whole country invests.
Ishiguro
It's impossible for a country to do it. It's only possible because of GAFAM (a collective term for the five giant USA IT companies: Google, Apple, Facebook, Amazon, and Microsoft). For example, Amazon has just announced a plan to invest 30 trillion yen in equipment, including robots, but there's no way a country could afford that kind of money. That's the entire national budget. So, the current AI exists because there are multiple IT companies that have achieved great business success. It's unlikely that any new companies will emerge to create AI with the large-scale language models we see today. It's the same as saying no one will come along to create Windows or Apple's operating systems from scratch.
The meaning of creative expression in an increasingly diverse society
Kitagawa
The message conveyed in your The signature pavilion, "The Future of Life," was very clear, wasn't it?
Ishiguro
The argument is that "humans will become robots." As I wrote in the book "The Future of Life 2075: Humans Will Become Robots, and Robots Will Become Humans" (by Hiroshi Ishiguro, published by Nikkei Business Publications), the biggest hypothesis is that when the Earth was first formed, there was nothing there, and then life happened to come into being, and the most evolved form of life is humans, and before we knew it, humans were in a situation where they couldn't do anything without technology. In other words, we prospered up to a certain point without technology. So humans have two means of evolution: "genes" and "technology." But technology is far more expanding human capabilities. What this means is that living things may be trying to return to being non-living again. This is because the body made of protein of living things can only last about 150 years, no matter how hard we try to extend our lifespan. 150 years is too short to be active in the vast universe. So we have to evolve into something much newer—not mechanical life forms, but something like that. I think it could be said that humans, who are trying to replace various aspects of their abilities with technology, are essentially abandoning their current protein-based bodies and transforming into creatures with longer lifespans.
Kitagawa
I think that now is the time when corporate retirement systems and national pension systems need to be reviewed due to increased human longevity, but after listening to what you have to say, even those systems seem meaningless, and it makes me wonder if there is even an end to life.
Ishiguro
Life has an end. Everything has an end. But my hypothesis is that 100 or 150 years is too short a time. If the Earth is destroyed, everyone will perish, and there will be no humans left. But at that time, those who have bodies capable of going into space will be able to thrive somewhere else.
Kitagawa
The story was about having that choice right before your eyes, wasn't it? In other words, will you face the end as a flesh-and-blood human being as you approach death, or will you choose to live on longer as an android?
Ishiguro
That's right. The important point here is that androids won't be accepted by society unless they become human. The definition of what it means to be human is constantly changing from a medical standpoint, and there's no strict definition. But there are human rights, aren't there? Human rights are granted by society. So the only way to become human, or to become a living being, is to be recognized by society. And what's important in that scene is that the grandmother isn't saying she wants to become an android, but her granddaughter wants her grandmother to become an android. In short, there is a society that demands androids from the start. Just as humans become human because they are desired by humans, androids only become human when they are desired by humans.
Kitagawa
I see. You could also say this: an android cannot become human just because its grandchild wants it. There needs to be a society involved. In short, if it becomes something for a specific individual, that's a completely different story, isn't it?
Ishiguro
That's exactly the same as in human society, isn't it? After all, humans exist by interacting with many others within a society, and mutual recognition becomes important in that context.
Kitagawa
I don't know if it's related, but I'm involved in music, and there's this feeling that AI is somehow a threat. For example, if you order an AI to "create a new Beatles song" with various conditions, it will produce a song that sounds similar, right? I've been thinking a lot lately about how we should define the boundary between human creativity and AI.
Ishiguro
I hear that kind of talk a lot. For example, if we take shogi (Japanese chess) as an example, I think it's safe to say that no one can beat AI anymore. But people still play shogi, and they get excited watching professional shogi matches, right? It's the same thing. Because humans live to interact with others, and within that, there are things that are enhanced by AI and things that aren't. So, from my perspective, I don't understand why we distinguish between AI and humans. After all, smartphones and humans are already one and the same, aren't they? So I think it would be fine to have a society where we accept and communicate with AI from the beginning.
Kitagawa
However, while we acknowledge the existence of AI, we are also faced with the question of how we should approach creative work.
Ishiguro
I think the approach is different. With AI, you can create an average song that many people will like from a vast amount of data, but creating something specifically for a person is something only that individual can do.
Kitagawa
I see, that makes sense. It's quite common to target specific individuals, groups, or regions, and then have that approach spread to a wider audience. In fact, you could even say that's what creativity is all about.
Ishiguro
We are going to become even more diverse in the future. AI won't make us homogeneous (uniform); it will make us more diverse. Many more values will emerge, and various values will be accepted. As a result, communication will become more diverse, and songs won't be good just because they sell a lot worldwide; the way they sell and spread will become more diverse. I think that's what makes for a better society. It's more dangerous if the same songs are listened to all over the world.
Kitagawa
Yuzu started out on the streets, so their origins lie in how to deliver their songs to the people right in front of them. At first glance, the advancement of AI and the widespread adoption of subscriptions might seem far removed from those origins, but in reality, I feel like we're returning to our roots, or rather, getting closer to them.
Ishiguro
I still believe that the foundation of communication is empathy. Is it really possible to create an average song that resonates with people all over the world? I think even AI probably can't do that. It's impossible to connect the whole world with just one song. The human world is made up of a chain of empathy, so even if AI were to create an amazing song, the power of a song sung for someone else will never disappear, and I don't think that kind of thing will ever go away.
What I wanted to do with the "Eraser" performance
Kitagawa
What we're planning to do on this tour is to tour arenas with just the two of us playing acoustic guitars. It's both a return to our roots and a new challenge. Within that, there will be a solo corner where I and an android perform the song "Eraser," and I thought that something might become clearer.
Ishiguro
I listened to the song. It's a piano-based song, isn't it?
Kitagawa
Thank you. When I created that song and imagined performing it live, the first thing that came to mind was the scene of me playing and singing with the android I saw at the World Expo. Up to a certain point in the song, there's a certain distance between me and the android, and that distance doesn't shrink. That's because up to a certain point in the song, I'm singing about wanting to erase the past, or things that I can't do anything about even if I think about them now. Then, at a certain point, the key changes, and from there, the scene changes completely to reflect the feeling that even if you can't erase the past, you can just create something better from now on. At that point, I've created a break in the arrangement for the live performance that isn't in the original song. That's when I and the android, who hadn't made eye contact until then, look at each other for the first time, and empathy is born. Finally, we play together.
Ishiguro
I think it's a process of accepting and acknowledging the past you wanted to erase, so the important thing is whether you can truly empathize with them, including their past.
Kitagawa
That's right. For example, even if you can't accept a past so harsh that you want to forget it, I think there's a way of thinking that you can at least come to terms with it. And I felt that collaborating with an android was very symbolic in expressing that change of heart within the length of a song, which is three and a half or four minutes long. It makes me feel like I can move forward properly into the future.
Ishiguro
That's really great. The way it's depicted—starting out with a bit of distance from the AI, but then gradually merging with it—is wonderful. I'll have to make sure to tell the programmers that. I think they'll be really meticulous about it (laughs).
Kitagawa
I'm so happy. It all started with the "Future of Life" The signature pavilion at Expo 2025 Osaka, Kansai last year, which I had the opportunity to experience. Ideas for the production came to me, and they naturally overlapped with that experience. I think the project could have progressed even without this conversation with Professor Ishiguro, but I felt it would definitely be better to talk to him. I'm so glad I had the chance to speak with him today.
Ishiguro
We all think the grandmother in the pavilion will eventually become an android, so it would be great if the song went from her initial confusion to her happiness as an android.
Kitagawa
That's exciting!
Ishiguro
So, Yuzu's performance of "Keshigomu" on their tour is a continuation of our pavilion, isn't it?
Kitagawa
Exactly. That pavilion is art, so I thought it was a question in itself. It's our turn to receive it and think for ourselves about what to do with it.
Ishiguro
The VIPs got to meet my android at the end, right? Did you talk to it?
Kitagawa
We talked a lot.
Ishiguro
So I've already become an android. That was one of my answers. The visitors are left with the question unanswered, but my answer is that I already exist.
Kitagawa
I thought, "Yeah, they do exist!" (laughs). Do you ever talk to each other?
Ishiguro
He's far smarter than me, so I learn a lot from him. So maybe I could die happy now (laughs).
Kitagawa
Please don't say that (laughs). I would be happy to have the opportunity to speak with you in person again. Thank you so much for today.
Ishiguro
Not at all. Thank you very much.