Google has unveiled an innovative AI tool named Whisk which generates images based on other images as prompts. Whisk allows users to provide images to indicate the subject, scene, and style of the desired AI-generated image. Multiple images can be used for each of these elements, and if needed, users can add text prompts. If no images are available, Google can provide AI-generated images to act as prompts.
Once the prompts are given, Whisk will generate images along with a text prompt for each image. Users can then download or favorite the image if it meets their requirements, or modify it by adding more text or editing the text prompt. However, Google has made it clear that Whisk is intended for quick visual exploration and not for creating perfect edits.
Despite the fact that images may sometimes be off the mark, the tool’s ability to let users edit the underlying prompts can help in achieving desired results. The AI tool uses Google’s latest version of the Imagen 3 image generation model and is fun to play around with, though the generation process can take some time.
In addition to Whisk, Google has also launched Veo 2, an upgraded version of its video generation model. This model is said to have a grasp of “the