Project Description

The prompt is a critical factor in generating text-to-image AI images. A prompt can be understood as an instruction. Yoko Ono’s book Grapefruit, first published in 1964, consists of instructions. They inspire the reader to have a variety of experiences—mentally, or even physically. They contain touchstones, like ‘home,’ or ‘wind,’ and while not lengthy or complex, they spark the reader’s imagination, creating what might be considered the actual art.

Her instructions have parallels to today’s AI prompts. Unlike AI prompts, however, Ono’s instructions spark the imagination of each person who reads them, triggering responses based upon their personal experiences. AI prompts, meanwhile, trigger algorithms to generate images resembling the prompts, but the results are significantly predetermined by the machine learning the AI has undergone prior to the user’s input. This machine learning is replete with the intentions and biases (known and unknown) of the person and/or institution conducting the process. These results are intended to satisfy a broad audience, drawing on visual touchstones (often clichés) that are commonplace in mainstream society.

Yoko Ono AI utilizes common, inexpensive AIs to generate images using the writing of various artists for prompts. The project started with Ono’s instructions in Grapefruit, but expanded to other well-known artists who have published writing. To date these include Sol LeWitt, Ai Weiwei, Karen Finley, John Baldessari, and most recently Marcel Duchamp.

The resulting images are overlaid with the artist’s text-turned-prompt, making it impossible to view the image without also reading the text. This visually and conceptually foregrounds the prompts—as well as their authors—and subverts the notion that the AI image is an autonomous, neutral entity that can be regarded as art.