“Virtual YouTuber” - the term refers to characters or avatars that post or release videos as YouTubers. Ever since Kizuna AI called herself such back in December 2016, numerous virtual YouTubers have appeared, thereby establishing the concept and recognition of virtual YouTubers. In this interview, we were able to talk with Takeshi Osaka, CEO and Founder of Activ8 Inc. which first proposed the idea of virtual YouTubers and is now the producer and in charge of the world’s most popular virtual YouTuber, Kizuna AI.
 

Could you tell us about virtual YouTubers first, for JAMLAB. readers who are not familiar with them?

First, the word “Virtual YouTuber” itself was coined when Kizuna AI started referring to herself as such in December 2016. Anime characters or CG models that are active on video platforms are called “Virtual YouTubers.” They are called “Virtual YouTubers” because they started on YouTube, but there are also CG characters who are active not only on YouTube but also on TikTok or Instagram; they are called virtual influencers as well. At present, there are more than 10,000 virtual YouTubers in Japan, and they often look like anime characters and often refer to them as such.

Virtual YouTubers, also called VTubers, come in many varieties. The first virtual YouTubers were mostly fantasy-driven; they were characters who act and move like humans as if the character itself is alive. From around 2018, the number of virtual YouTubers started to increase all of a sudden, but the reason for this was not because of the richer looks or movements of 3D models, but because of the increase in virtual streamers that use 2D illustrations of characters moving in a simpler way. Characters of virtual streamers do not move much, but people enjoy their talk and their personality. I think there are more virtual streamers now. One characteristic of virtual streamers is that they can be run by individuals and can be done more easily.
 

© Kizuna AI

Thank you very much. I think it’s easier to understand if we compare it with Hatsune Miku. Would you say being able to interact with virtual YouTubers in real-time distinguishes them from Hatsune Miku, whose conversations and songs are made using Vocaloid?

You can say that. Being able to communicate with them is the biggest difference after all. In Hatsune Miku’s case, I think the main point is that creators make content and it is strong in the aspect of using voice software to produce music, while in the case of virtual YouTubers, the character is the main actor so content released by the character itself makes it different. Hence, with virtual YouTubers, there’s the story, context, and values of the virtual YouTubers themselves that can be applied in life. I think that virtual YouTubers have evolved to provide value that is relevant to current times by continuously providing interactive entertainment.


When Hatsune Miku appeared, it impacted the industry so much that many people entered the Vocaloid business. Likewise, virtual YouTubers are rapidly increasing in number, so could you tell us what the current situation is?

Let’s see...there are different types of virtual YouTubers, as I mentioned earlier, some mainly work as (virtual) streamers, and others are active on TV and the new platform like TikTok, publishing manga, or doing music. The range of their activities is expanding in all aspects, similar to human beings.


So currently, they start as something similar to YouTuber, then expand to other activities.

Yes. I think it's really becoming diversified as a business. In the case of Kizuna AI, I think it has diversified considerably in terms of both business and content, such as appearing in TV commercials, appearing at music festivals, and developing merchandise.


Moving on to Kizuna AI who currently streams her gameplay, could you talk about the process and maybe some inside stories that led to her birth which you can share with us?

Before founding the company, I worked closely with the anime industry. I really felt the power of IP (intellectual property) in Japan, and at the same time, as technology has evolved and platforms have changed from TV to YouTube, I was questioningabout how anime content should be. The animation industry is gradually evolving, but if you look at it from a broad perspective, the way it is done hasn't changed much, but on the hand, I felt it’s potential. If we use technology, characters will be able to step out of the world of anime and become influencers and act like humans. There was a time when I was talking with my friend about how amazing it would be if such a thing were realized. It was at this time that I met Kizuna AI. She inspired me and I said, "This is it!" and Project A.I. started.

There was no virtual YouTuber at the time, so we combined "business development in today's era," "the feeling of excitement with futuristic feels," "maximizing the potential of anime characters," and "taking advantage of Japan's strengths" and decided to use the platform YouTube. At the time, many YouTubers were releasing content daily, so it's important to do the same to maximize YouTube’s potential. We started from the point that releasing content every day was not doable because of the production period and costs, but we improved the production system and content. I also thought it had to be popular overseas, so I decided on the image of the character's visual and voice one by one. Naturally, we worked gaining the cooperation of various people, including creators, illustrators, and engineers in the process.


What do you mean by deciding on the character visuals that would be popular overseas?

There are pros and cons because how they are seen and perceived differs from person to person, but first of all, Japanese anime characters often look young and has small body proportion. From the point of view of overseas fans, it can be said that this is just a tiny difference, but in the case of Kizuna AI, we put importance on being easily accepted by many people, taking various factors into consideration such as making her body proportion taller on purpose and making sure that she does not look unpleasant as a 3D model.


Indeed, it would be awkward if the characters are too realistic (laughs).

We did it like so because we couldn't maximize the strength of Japanese anime content. There is a game called "Final Fantasy" and the recent "Final Fantasy" games look very real. I originally wanted that taste of reality. However, the co-producer said, "If it is too realistic, it cannot maximize the Japanese touch and strength of Japan, so it should be more like an anime character." We had a heated discussion for several days. As a result, the visual that illustrator En Morikura came up with is close to the ideal. That’s one inside story for your readers (laughs).


When you see music videos of Kizuna AI, I think the graphics are amazing. If you look back in time, there was a sensational animation like Love Live!. The animation was created from the motion capture of dancers, and that naturally took a lot of money and time, but I was surprised at the quality at the time. Four years have passed since then and Kizuna AI has appeared, and I feel that something even more amazing has appeared. Could you tell us a little about the technology?

Actually, Kizuna AI is an AI (artificial intelligence), so I can't tell much, but in terms of the technology that many other virtual YouTubers use, the fact that low cost, easy and accurate motion capture has come out and programs to reproduce such can be used on a low budget are huge factors. The improvement and convenience of technology to make characters move in real-time also contributed greatly.


Does it mean based on the evolution of technology, it was optimized for virtual YouTubers?

We are not the ones who make real-time animation in the form of live animation, but to some degree, it was already being done on TV. The technology had been out there for a while, and it wasn't very difficult in technical terms, but going beyond the boundaries of animation and being used as a YouTuber is something new. Technically, the technology has been used in Hollywood for a long time, and it became very accessible to common people. I think connecting these technologies and business models is what’s important. Without this connection, we won’t be able to sustain it and we can't establish it as a new genre of content. Viewers see us as YouTubers, living and doing business like humans. It's important to refresh users’ perceptions. One of the reasons virtual YouTubers have increased is that the technology required for characters to act like YouTubers on YouTube has become commonplace.


Kizuna AI’s activities are diverse. What made her start her music career?

The catch copy and mission of Kizuna AI is "I want to connect with everyone." "Everyone" refers to every single person in the world. Of course, music content can expand across national borders, so harnessing that power was one thing. Being a YouTuber comes next and since she can communicate with others, I think offline events are very important for business. I think the value of real events has relatively increased in the music industry. I think when you organize an event of something, a music event is the best thing one can enjoy unconditionally. When you increase the value of user experience, you could function as the hub of the world; these are the two reasons why she started her music career.
 

I agree that music has the power to go beyond borders. If you are taking foreign countries into consideration, you are certainly on point.

As for the genre of music, I produce dance music because I think it’s something that can be fully enjoyed and that it will be recognized by more people. I love Japanese anisong and pop songs, but EDM has become very popular around the world, and I thought dance music would mission-wise, enhance Kizuna AI's potential. Fans also don't move their bodies when they watch videos, but during events, we would like them to move and jam along as much as they want. With that in mind, we focus on dance music.


So you work with creators who are skilled at dance music to produce her music.

Yes, we did collaborations with TeddyLoid, Yasutaka Nakata, ☆Taku Takahashi and DÉ DÉ MOUSE.


Music-related videos tend to be viewed more often. Do you find the business of having the rights to the music and earning from it appealing?

I am still learning about the music business and I have high expectations, but as I mentioned earlier, our marketing side is relatively stronger. As you pointed out, the number of views of music content is increasing, and I am happy to say that the number of views of original music is particularly high. I think we should increase awareness of our brand first and if we can do business from there, that would be great.


The more songs played, the more royalties you get.

We use many different types of music and background music (BGM) when we make videos and hold events, so naturally, having music as the content is a must. Music is highly content-based, and it's a very valuable genre because it's played and watched repeatedly.


Next : Takeshi Osaka Interview Part.2 will be published on 2 March, 2020