3DALL-E Integration of Text-to-Image AI in CAD Workflows

This document is a research article that discusses the integration of text-to-image AI, specifically DALL-E, within 3D design workflows through a plugin called 3DALL-E. The objective is to explore how AI-generated image inspiration can assist designers in creating 3D models using CAD software. The authors describe the functionality of 3DALL-E, which generates 2D image inspirations based on user-defined text and image prompts, enhancing the design process. A user study involving thirteen designers is detailed, revealing that participants recognized the potential of 3DALL-E to aid in conceptualization, prevent design fixation, and inspire new design ideas. The article also discusses prompting patterns observed during the study and proposes the concept of prompt bibliographies to document the interaction between designers and AI. The findings suggest that integrating text-to-image AI can significantly benefit CAD workflows by providing relevant design language and visual references.