Forget prompts! Google's Whisk lets you create images with images
Google has unveiled a new artificial intelligence (AI) tool, Whisk, that lets you generate images using existing ones as prompts. This innovative approach sets itself apart from traditional methods that rely on detailed text descriptions. The company calls it "a new type of creative tool" designed for experimentation and innovation. Whisk is part of Google's latest generative AI experiment and is currently accessible through Google Labs in the US.
Whisk's functioning and user experience
Whisk works on Google's Imagen 3 image generation model, a new development announced with the tool. It is meant for "rapid visual exploration" instead of precise pixel-level edits. Google has clarified Whisk may not always yield expected results, but users have the option to edit underlying prompts for refinement. The tool is especially useful for creative professionals and artists who want to quickly explore ideas.
Whisk's unique features and customization options
Whisk offers a simple interface with pre-defined styles like sticker, enamel pin, and plushie. It also has an advanced editor mode for further customization. The tool captures the "essence" of the input image instead of creating an exact replica. This may lead to variations in characteristics like height or skin tone, giving a unique touch to every image generated. Google has emphasized that Whisk's outputs are more appropriate for brainstorming and quick visualizations than production-ready content.
