Google DeepMind

Art&Graft invited me to join their design team and explore the mood of some key moments of a motion video for Google DeepMind. The goal of this motion piece was to present MuZero, an Artificial Intelligence (AI) technology using machine learning to learn games by itself and then fully master it.

Here are some WIP styleframes from early concepts that didn’t make it to the final motion video:

↓

We started by exploring abstract and color-driven imagery but also some more illustrative and recognizable shots with video game controllers, chess pieces, and so on.

The greatest challenge of this project was to jump from a colorful and abstract shot with typography to an illustrative chess board or a Breakout Atari game. The main element that glued everything together was definitely the color and the sense of an infinite and borderless space where everything happens.

These next styleframes were animated and are part of the final motion piece:

↓