Google has introduced a new AI image generation tool called Whisk that enables users to create images using existing images as prompts rather than text descriptions. According to Jay Peters from The Verge, the tool allows users to specify subject, scene, and style through image inputs or Google-suggested examples. Users can refine generated images through additional text prompts or by editing the underlying parameters. The system runs on Google’s latest Imagen 3 model and produces results within seconds. While Google acknowledges that Whisk is designed for quick visual exploration rather than precise editing, it offers features like image favoriting and downloading.