Hello, I’m Kai.

* * 　　　 ⊹ ˚ . 　　. ⊹ · ✧ 　　　* * . 　　 ⊹ ✦ 　 ✵ 　　　　　* * .　 · . 　 　 * . ˚

I’m Kai, nice to meet you. I grew up in Okinawa, Japan, but now live in Los Angeles. I graduated from USC’s School of Cinematic Arts with a degree in Film and TV Production. As an animator, my intention is to work with machines to copy nature. I enjoy finding technology to incorporate into film and animation. This led me to work with machine learning tools to combine live-action and animation. I started a little company called ARTI in 2022. We’ve made projects in IMAX and VR and recently finished interactive concert visuals. I’ve provided work for professional clients:

I love the Avalanches, a musical group with a huge library of samples that they collage into songs. I get the same joy by bring- ing video samples to life. My aim is to use sampling of all types to make animation. I’m currently working on a music video for Warner Media and will be working with music labels to create a music-driven feature-length horror. Irrelevant, but my mom is a florist, and my dad is an avid meditator.

I’ve worked with A.I. tools since 2019 when NVIDIA came out with their GauGAN. It was a simple but mind-boggling tool at the time that allowed people to paint landscapes. I used that to make my university’s first A.I. animation. When COVID hit, my curiosity about machine tools kept growing. Over the course of the last four years, A.I. tools kept getting bigger and better—and I tried my hardest to keep up. I’ve experimented with face swapping, voice cloning, neural radiance fields (3D landscape stuff), facial augmentation, body morphs, motion capture, style transfers, machine music, generative imagery/video, text-to-image, text-to-video, image-to-video, video-to-video, and more. There’s a lot out there.

A few years ago, A.I. was a tough hobby to start. Most machines aren’t strong enough and services like ML runway weren’t around. So I ended up adapting these tools for Google Colab—a cloud computing site. Only to discover people much smarter than I have already done so. There was a mountain of open-source A.I. tools put together by a community and made available for free. Honoring that, I avoid private/corporate models. A goal of mine is to only incorporate open-source models and to use A.I. not in a way that avoids work but relies on it. Whatever you define artistic cheating, A.I. has made it easier to do so. I’d like to avoid doing that. I’ve worked to incorporate A.I. into private and corporate workflows. In my experience, each workflow reflects the values of the individual or the company.

I got into A.I. tools to work with artists in different mediums. I lived in a little house of painters, composers, and actors. These tools were a way to bridge that gap—to bring live actors into animation, and stylize film as paintings. This is what makes A.I. so fun for me.

The growth of A.I. may be exciting, but there’s still a great need for good, high-quality A.I. content. We need art that relies less on novelty but rather focuses on new avenues of working. And spending time making something good.

I want to sit at a desk and work for the content I make. I don’t wantto give that up to a machine or an automated pipeline. I’d love to work with a team to take a step toward that reality. And in the process, expand my own experiences with new types of tools and collaborators.