Google has launched an AI experiment called Whisk for Labs testers, which allows users to generate images by uploading images that define subject, scene, and style. Whisk uses Gemini and Imagen 3 to extract characteristics from the uploaded images, creating variations in attributes like height, hairstyle, and skin tone. It features a "review and edit" option for users to refine generated images and provides a detailed description of the creation process. As of December 16, U.S. Labs testers can sign up to explore Whisk.
Additionally, Google has updated Imagen 3 and Veo 2. Imagen 3 now offers brighter and better-composed images with improved adherence to prompts, rolling out globally in ImageFX within Google Labs. Veo 2 has been enhanced to create high-quality videos in 4K resolution, allowing for detailed descriptions and improved understanding of expressions and movements. The updates for Veo 2 are being rolled out in Google Labs' VideoFX, with plans to integrate it into YouTube Shorts and other products in the future.