AI artist performance combined with chef dinner

•

Dinner-performance Pacific Diffusion is a presentation of a project at the verge of gastronomy and contemporary art, created in a collaboration between a chef and an AI artist.

In the gastronomic and art set Pacific Diffusion, brand chef Egor Anisimov and AI artist Vyacheslav Furashov explored the topic of borders – as lines of contact between different worlds. Lands and seas. East and West. Tradition and the future. Human and artificial intelligence. Gastronomy and art.

Primorsky Krai is the point where the borders of Korea, China, Japan and Russia converge. A place with an ancient and rich history. We wanted to reflect this connection in our work.

I used artificial intelligence to visualize the descriptions of the chef's dishes. Stable Diffusion 1.4, with my help, saw this gastronomic set as a series of unusual paintings: distinctly Asian, but with a European accent; as a combination of soft photorealism and the styles of contemporary artists with classical engraving. We can say that this is a gastronomic review that was made by an unreal person.

I started with writing down the words and images that Egor and I came up with when creating these dishes. Then I processed the received data to get a more stable and predictable result from the AI. In the process of this work, I came to understand that artists and artistic styles from countries whose culture has influenced our region can be used to explore the topic of «borders». So in the works there were references to ukiyo-e, nihonga, manga, traditional landscape paintings of Korea and China.

The original images I was able to get with Stable Diffusion were low resolution, only 896 pixels long. So I used the Gigapixel upscaler to get files large enough to print in A1 format. In this case, the appearance of artifacts in the image is inevitable, but I managed to fix this by printing on a thick canvas on which these defects were not visible.

During the course of my work, a Google.Colab notebook that allowed me to make videos from a standard prompt for Stable Diffusion came out. So I immediately took advantage of this opportunity – 7 small videos were added to the project, which illustrated the process of AI working on each of the pictures. They were projected on the screen during the performance.

I also considered the possibility of using AI generated music for each of the pictures, but with the Сollab notebooks options available at that time, I could not achieve the desired results. So we limited ourselves to samples with the sounds of the sea and the forest.

Before serving the new dish the clothes of the waiters were sprayed with wet paint, the amount of which increased as the performance rolled on. At the end of the performance, these shirts were also turned into works of art.