Lat month, Google announced that Duet AI would be coming to Workspace, and was available to preorder for Workspace Enterprise customers looking to use generative AI to write, visualize, organize, and accelerate workflows.
At the time, the free office software provider confirmed that Duet AI for Workspace would be generally available “later this year.”
However a 9To5Google report now claims the Duet AI text-to-image generation tool has expanded to more accounts as the company continues its public testing of the Imagen-powered function.
Google Workspace text-to-image AI
A new pop-up that reads “Help me visualize” serves as the key user interface for the tool, which Google says is powered by a number of image-generation models including Imagen.
Users can interact with the text box, which allows them to type any combination of commands that include the desired image’s contents, colors, and styles. Duet AI will then make eight results available to select for use.
There are also pre-determined styles to choose from: none, Photography, Background, Vector art, Sketch, Watercolor, Cyberpunk, and I’m Feeling Lucky.
Currently, image generation is not instantaneous and takes around 20 seconds, so it’s longer than the company’s AI writer. Already, Google’s Bard feels a touch slower than its key rival, ChatGPT.
Google is currently testing the tool in Slides, and those who have access to it can navigate to slides.new where they will see a pop-up the first time it becomes available.
Google says that, in trade for early access, human reviewers may read, annotate, and process interactions. As such, customers are advised never to include personal, confidential, or sensitive information when using Labs features.
Source link
