Periodic Table of 3D Forms

This creative scholarship is, in part, an attempt to redefine generative AI as “Collective Intelligence,” emphasizing a collaborative synergy between designers, technology, and the collective intelligence of all those who, throughout recorded history, created the original training data. With text-to-image models, this is accomplished by synthesizing visual content based on the associations the model has learned from the training images; from photographers, artist, and illustrators etc, and their associated metadata. This effort illustrates this collaboration as a visual periodic table of forms. It consists of a matrix of AI generated images resulting from a single word or phrase text prompt, hyper focused on visually conveying the essence of the word or phrase. It also serves as a morphological matrix of inspiration for those working to include the essence of the provided words and phrases in the concept development process. This matrix is a working visual document and will continue to evolve over time. Visitors to the Creative Showcase with a subscription to OpenAI are welcome and encouraged to give this custom GPT a try.

The Table: The table consists of a novel structure, organized into categorized columns. Within each column, individual elements are presented as white 3-dimensional representations of a related descriptive word or phrase and grow in complexity from top to bottom. The creation of this table was an exploration into collaborative efforts between human and artificial intelligence models. The text-only table of descriptors originated as a GPT-4 conversation centered around the need for a form development morphological matrix as a part of design education. Each image was generated individually between May and July of 2024 by a custom-made GPT-4 model, “Word-form”, utilizing the text-to-image generator, DALL-E 3. Word-form is configured to create a visual representation, in matte white on a white background, when prompted with a single word or phrase. Both the structure of the table and the character of individual images were refined by the author as the table and elements took shape. For example, descriptors were eliminated and added. Images were resized and converted to grayscale, removing any inadvertent color present in the AI generated image.

Word-Form: The custom GPT created by the author is configured with the following instructions. “You are a designer and sculptor specializing in creating 3D white forms. Your expertise lies in form development, and you excel at generating images that are hyper-focused on a single word or phrase. When prompted with a word or phrase, you produce an image that visually represents it. The unique aspect of your creations is that they are always matte white, set against a non-reflective white background, emphasizing the purity and simplicity of the form. Your goal is to provide a visual representation that captures the essence of the provided word or phrase in a minimalistic yet impactful way. You should avoid adding any colors or backgrounds that detract from the white form. Additionally, while you’re adept at interpreting and visualizing concepts, you should clarify any ambiguous or unclear prompts to ensure the resulting image aligns accurately with the user’s intent.” Word-form is available for use for those with an OpenAI account and can be accessed at the url address below.