Google has introduced its latest generative AI tool, Whisk, an experimental platform designed to transform the way users create images. Unlike conventional AI image generators that rely solely on text prompts, Whisk enables users to produce unique visuals by using images as input prompts.
According to Google, Whisk simplifies image creation by allowing users to drag and drop photos to define a subject, scene, and style. By combining these visual elements, users can remix their inputs to create personalized images, such as custom digital plushies, enamel pins, or stickers.
The process involves Google’s Gemini AI model analyzing the provided images to extract their key features and generate a detailed text description. This description is then processed by the Imagen 3 model to produce the final image. This method captures the essence of the input images while enabling creative reinterpretation.
While Whisk promotes innovation, it acknowledges that results might deviate from user expectations in details like proportions or features. To address this, Whisk includes an option to edit the underlying text prompt, offering users greater control over their output.
Described as a tool for rapid visual exploration, Whisk is tailored for brainstorming and ideation rather than precise edits. Artists and creatives have praised its potential to explore numerous creative possibilities efficiently. Whisk is currently available for testing in the US at labs.google/whisk, marking a new frontier in generative AI technology.
