Text-to-Image Generation

An artificial intelligence capability that creates visual images from natural language text descriptions, also known as prompts. Tools such as DALL-E, MidJourney, and Stable Diffusion use large-scale diffusion models trained on image-text pairs to generate novel images matching user specifications. In assistive technology design, text-to-image generation can help makers visualize AT concepts without drawing or CAD skills, serving as the first stage of an AI-fabrication pipeline. However, these tools often struggle with specialized AT terminology, produce generic rather than disability-specific designs, and may reinforce biases present in training data that underrepresents people with disabilities.