Yumi Matsutoya, known affectionately as “Yuming,” has never stopped looking ahead. Over a 54-year career that has made her one of Japan’s most beloved musicians — with more than 42 million records sold and a treasured place in the Studio Ghibli soundtrack canon — she has treated technological shifts not as a threats but as invitations.
In the early 1980s, she was one of the first major Japanese pop stars to experiment with digital synthesizers, folding their tones into her songwriting as naturally as if they had always belonged there.
Her 40th studio album, “Wormhole,” finds the 71-year-old artist testing the latest potentially epoch-defining technology: artificial intelligence. Her interest in AI began a few years ago, when researchers at the University of Tokyo approached her with a proposal to reconstruct her voice as it sounded in the early 1970s, when she debuted under the name Yumi Arai. Working with previously unreleased vocal takes from that period, she created “Call Me Back” in 2022, a song that plays like a quiet duet between present and past. It was one of the first songs to prominently feature AI-assisted vocals, and helped spark the concept for “Wormhole,” which arrives at a moment when the technology is both commonplace and fiercely debated.
Matsutoya has avoided the generative tools of this trade — ChatGPT and music-making programs like Suno — that are driving much of the controversy. Instead, she and her longtime producer (and husband), Masataka Matsutoya, turned to Synthesizer V, an AI-assisted singing-synthesizer, to merge her present-day voice with its Arai-era counterpart. The result is what they call a “third voice,” named “Yumi AraI,” a spectral presence that threads through the album.
“Wormhole” doesn’t use AI as a crutch. Rather, it becomes another tool in Matsutoya’s artistic arsenal. The songs find her crafting the kind of mid-tempo pop-rock from her early career, fusing it with elements of progressive rock and opera — human-made first and foremost. On her current nationwide tour, the AI voice appears alongside her onstage, woven into elaborate sets and dramatic costumes, the digital at ease with the practical.
Several weeks before the album’s release on Nov. 18, Matsutoya sat down with The Japan Times for her first-ever interview with an English-language publication. Over the course of an hour, she spoke about “Wormhole,” city pop, AI and how artists can approach new technology on their own terms.
Are you the kind of person who looks back on the past a lot?
I’d say I look back on the past about the same as I think about the future. In my mind, the past and future exist in parallel.
What, for you, is most exciting about the future?
It would be technology. It has two sides, a good and bad one, but if I focus on the positive, it gives me a lot of hope. I won’t be around in a hundred years, but imagining what kind of culture will be happening and popular then really excites me.
That feels like a good jumping-off point to talk about your 40th original album, “Wormhole.” When did you first encounter artificial intelligence as a technology?
In 2022, a team at the University of Tokyo approached me with the idea of reconstructing my voice using AI technology. I gave them old vocal tracks, especially ones from when I (recorded as) Yumi Arai, up through 1976, and they created an “AI Yumi Arai voice” from the vocal tracks. Then I made a song (“Call Me Back”) where the old Yumi Arai and the current Yumi Matsutoya — me today — “duet.” I even sang it on NHK’s “Kohaku Uta Gassen” in 2022.
What made you want to pursue this idea they brought to you? What about it made you think, “This sounds cool, I want to try it”?
Pure curiosity. I’m proud of my long career. It’s not a challenge many artists can attempt, but someone with a long career like mine can.
You’ve always seemed curious about incorporating new technology into your music — from new recording processes to being among the first pop acts in Japan to use the Synclavier digital synthesizer. Why do you think that is?
Both my producer — my husband, Masataka Matsutoya — and I are extremely curious. We complement one another, and together it gives us a turbo boost [laughs].
Once you decided to make “Call Me Back,” what was the process like?
First, we found an unreleased track buried deep in the Yuming vault. Actually, the working title of that song was already “Call Me Back.” At the same time we came across it, the University of Tokyo connected with us. It felt like a perfect match for this experiment.
What did you learn from making “Call Me Back”?
That people change over 50 years … even their voices. [laughs] I don’t think my look has changed much, but still, it’s different.
What do you feel when you listen to the Yumi Arai period of your career? I imagine you’ve had to hear it a lot over the past five years, working on this project.
I sang very straight back then, almost flat, with no vibrato, but with lots of overtones, and my key range was higher. Later, through vocal training, I realized that when you’re younger, your chest voice sounds like falsetto, your chest voice is actually singing with a falsetto technique. And vice versa. It’s seamless. As you get older, they separate: chest sounds like chest, falsetto like falsetto.
When did you start working with a vocal coach?
About 30 years ago, I realized I needed training to make those registers seamlessly.
So you make “Call Me Back” and perform it on “Kohaku” in 2022. Then there’s “Wormhole.” What about that initial experience with “Call Me Back” made you think that you wanted to continue to explore the idea of AI-assisted track production and, eventually, build a whole album around it?
It was actually my producer’s idea. At first, I wasn’t sure whether I could actually execute it. But, curiosity won out.
What were your biggest doubts?
First there were worries over the transition from studio album to live performance. You have to play the song live, so would it even be possible to reproduce it on stage? Second, was the title and theme “Wormhole.” I worried it might be too difficult a concept to explain, that it might be too artistic for a pop album. This is a feeling I’ve experienced before.
When did you feel that before?
It has happened quite a lot. [laughs] I don’t always say it, but it crosses my mind a lot.
Fair enough, but isn’t that what artists do?
[laughs] That’s true.
I’ve personally experienced what I call “wormhole” moments. I do think most people experience them — our consciousness travels: memories, images, dreams, nightmares. These are “wormhole” phenomena to me.
Yumi Matsutoya’s 40th original album, “Wormhole,” sees the artist singing alongside past vocals and an AI-assisted “third voice” that meshes the two.
How would you describe the album to someone first encountering it?
“It feels familiar, but you’ve never heard it before.” People don’t often listen to albums straight through these days, but if they could listen to “Wormhole” all the way through in one go, I think they will feel like they are watching a very beautiful science fiction film.
Are you a science fiction fan?
Absolutely.
What are some of your favorite science fiction works?
Ridley Scott is one of my favorites, as is Christopher Nolan — I’ve watched “Interstellar” so many times over the years.
Getting into the construction of the album, the hook is the use of AI — specifically, the Synthesizer V by Dreamtonics software — to create a “third voice.” What was that process like? I imagine it took a lot of time.
The producer spent the longest time on that. It was like weaving a tapestry. For me, because one of the voices would be utilizing AI, I could write songs without worrying about the vocal range or the tempo. I felt free to create the music I wanted to, and I thought, “I can train later to catch up.”
Is there a song on “Wormhole” that illustrates that freedom, one that allowed you to attempt something you might not have done otherwise?
(The sixth track,) “Karasu Ageha.” It has many key changes, and eventually returns to its original key. Without AI, the vocal harmonies would have been impossible to make. There’s also the last song, “Soshite Dare Mo Inaku Natta.” I wanted to write an aria, like an opera. I wanted it to be something like “Nessun dorma” from “Turandot,” a song I couldn’t sing with my current vocal delivery. With the help of AI, I thought I could create a song like this, and I trained using it. Eventually, I regained a range I hadn’t been able to use in a long time.
It’s like how, when shogi players compete against AI programs, their skills develop and improve. It was the same thing with me.
You’ve emphasized that while these new voices were created with help from Synthesizer V, you didn’t use generative AI for the music or lyrics. Why was that?
I never considered AI when it came to composing music. Writing songs is my artistic identity. I did try using ChatGPT to write lyrics, though. AI is … it needs a lot more work to be me. [laughs] Maybe it will catch up soon, but only a human can truly move another human’s heart. That’s something I learned during this process. Maybe ChatGPT can help you with a school report or business proposal, but in creative fields where there’s no single “answer,” AI can’t get it right. Only humans can make something from nothing.
There’s a lot of anxiety around the intersection of AI and art. It’s reassuring to hear you stress the need for human presence.
Maybe it’s the Japanese sense or the language itself that helps here. Japanese words carry a lot of meaning, and there’s a feeling of fuzziness and reading between the lines with them. There’s so much information contained in that space that AI can’t learn.
That’s why we get criticized by a lot of people outside Japan — the Japanese tend to be very “fuzzy” in how we talk. [laughs] There’s sometimes no subject in our sentences, but we usually understand what’s happening. For example, Matsuo Basho’s (17th-century) haiku, “An ancient pond / a frog jumps in / sound of water” really can’t be translated into English without losing that quality. Along with that space, the rhythm of Japanese words creates a qualia. That feeling like when you look at a sunset, and feel sentimental, or smelling the rain. That feeling. AI can’t capture that either.
Even with AI central to the process of creating “Wormhole” then, you’re still writing the music and lyrics. It’s still very much a human record. What was on your mind as you wrote?
The central lyrical concept is captured in the second song, “Cinnamon.” It’s the me of now and the “me” of the past communicating with one another. Every song contains memories or imagines another dimension. There’s a lot of “what if?”
On “Karasu Ageha,” I wrote that as the theme for the adaptation of a famous horror manga. It’s a song connecting the spirit world and this world, going back and forth between them. Coincidentally, I’ve been watching the NHK drama inspired by Lafcadio Hearn and his wife, which centers around ghost stories. Japanese ghost stories can be beautiful and nostalgic, not just scary.
What else did you want to explore musically?
On each song, I tried something I had almost done before but hadn’t truly done. [laughs] For example, the first track, “Dark Moon,” I wrote it as progressive rock, like Pink Floyd. [laughs] That was a challenge.
Given the concept of the album, if you could go back and speak to Yumi Arai at the start of her — your — career, what would you say?
I touch on that in “Cinnamon.” My past self asks, “Where am I now? Where did I go?” and my current self replies, “I’m here.” Both are me. I can’t answer more concretely because, inside, I haven’t changed a bit from that time. [laughs] I call myself a “super junior high schooler.” I’m always in this teenage, adolescence mindset [laughs]. Around 20 years ago, I experienced menopause, and I saw it as a positive. It was like I’d returned to my youth.
Because I started this work at 14 and just kept going, I’ve been able to protect that 14-year-old self. It’s my moratorium. I wanted to live listening to my favorite music and wearing the fashion I like. That was my dream and that’s how it turned out.
Since we’re talking about the past, what has felt to you like recent history’s most technologically advanced music development, the equivalent of AI today?
At the start of the 1980s, it was the Walkman. It was liberating! You could take it out to the poolside and listen to music in good quality. It seems like every 20 years something technologically transformative happens. In the 2000s, computer graphics felt revolutionary. Now AI in the 2020s. The iPhone might have been a preview of that, putting a computer in everyone’s hand to prepare them for AI.
Moving beyond the album for a moment, Japanese music from the 1970s and ’80s has become popular overseas in recent years. Have you been following that resurgence?
I haven’t followed everything surrounding it, but I can confidently say I’m the beginning of it, musically. [laughs] What’s now called “city pop” was heavily influenced by what I was doing.
I think that’s a fair thing to say. For people who only know Tokyo through city pop and have created their own imagined nostalgia of it, what was Tokyo actually like in the ’70s and ’80s?
From the ’70s to early ’80s, there was a feeling in the air that something new was about to begin in Tokyo, something fun. By the late ’80s, with the bubble economy, the sense of time shifted. But up to the early ’80s it was full of that anticipation.
Just out of curiosity — since I’m speaking with someone who helped create it — did anyone actually call it “city pop” at the time ?
No. Everything is probably kind of like that. You need some time and distance to get perspective on it. It’s like bossa nova in the late ’50s, when stylish kids in Rio de Janeiro made that music. It’s similar to what happened in the ’70s, all the stylish kids from private schools in Tokyo started making the sound and the scene — it was such a small world — and kids outside the city longed for it and thought, “That’s Tokyo.” And the movement grew.
You mentioned being worried at first about presenting “Wormhole” live. How are you approaching that now?
Only a week ago did I feel like everything clicked while rehearsing. It felt like we could translate it to the stage. Of course, we can’t recreate it 100%, it’s impossible. We’d need a way to freely control a hologram. If that comes about, maybe it’s possible. Still, I felt the way we combined analog and digital technology creates enough of an illusion.
Whether it’s facing AI or staging a show, it’s ultimately a means to reach people’s hearts with my songs. That’s why I’m facing this reality.
The tracks on “Wormhole” revisit the mid-tempo pop-rock of her early career, with touches of progressive rock and opera mixed in.
What advice would you give today’s artists about using AI or other technology for music or live performances?
Have a strong sense of self, both physically and mentally. If you know who you are, you can use technology efficiently to refine yourself. But you must have a strong core, especially in a future where a “copy” of you might be possible. You have to know who you are.
Also, hold fast to who you are. In the streaming era, a song someone discovers today is “new” to them regardless of trends. Don’t chase trends. Play counts can be surprisingly big, but they’re just numbers. Build a community that truly understands you.
How do you keep your own core strong?
By being proud of everything I’ve been doing over such a long time — all the experiences I’ve accumulated, musically. Confidence in doing this for so long, and all the experiences I’ve accumulated.
What are the biggest lessons you learned in making “Wormhole”?
Facing AI head-on taught me a lot. It made me ask myself, “Who am I? What is time?” I confronted many philosophical ideas I hadn’t thought about before.
Finally, is there anything you’d like listeners to keep in mind when they listen to “Wormhole”?
The message of this album is “live strongly.” That came to me during the shoot for the “Dark Moon” video. That energy reached me, and I put it into words. There are so many disasters and difficulties in the world, but you really need to keep a strong sense of self. That’s most important.
AloJapan.com