Google has announced the rollout of AI image generation for Google Slides, a part of its Duet AI initiative. This feature is set to change the way presentations are created, offering users a variety of styles to choose from, including photography, illustration, flat lay, background, and clipart.
AI Image Generation in Google Slides
The AI image generation feature in Slides, aptly named ‘Help me visualize', can create a variety of illustrations based on a text prompt in mere seconds. Users can select a specific art style with the prompt to generate images of a certain kind. The AI then loads a collection of visuals which the user can embed in their presentation. If the user doesn't prefer any of these images, they can click on ‘View more' to generate an alternative set of images for the same prompt.
How to Use the ‘Help me visualize' Feature
To use the ‘Help me visualize' feature, users need to open a presentation on Google Slides, click the slide where they want to add the image, and then choose the ‘Help me visualize' option. They can then enter a prompt, such as “A beautiful photograph of a Hawaiian beach at sunset with palm trees in the distance” or “A plate of freshly baked blueberry muffins on a wooden kitchen table next to a large window”. For better results, users are advised to include the subject, setting, distance to subject, materials, or background in their prompt.
Expansion of AI Capabilities in Gmail and Docs
In addition to the ‘Help me visualize' feature in Slides, Google is also set to expand existing AI capabilities in Gmail and Google Docs. The ‘Help me write' AI feature in Gmail, which can prepare email drafts based on simple text prompts, will now be able to suggest context-based replies. It will use the previous messages received in the mail thread to form context-specific reply suggestions. Similarly, ‘Help me write' in Docs will now add smart chips and other variables in the body of an AI-generated draft, helping users identify, for instance, the places where they have to insert the company name in a job application.
Google Workspace Labs and User Privacy
The new AI features are part of Google Workspace Labs, a trusted tester program for users to try new AI features. Google uses and stores data such as prompts entered or selected, image styles selected, generated images, and user feedback on generated images. However, users are advised not to include personal, confidential, or sensitive information in their interactions with Workspace Labs features.
Latest Developments in AI Image Creation
Two of the currently most popular AI image generators are DALL-E 2 and MidJourney. DALL-E 2 was created by OpenAI and can produce realistic and original images from natural language descriptions. MidJourney is an AI image community that allows users to generate high-quality images for various purposes and platforms.
In March, Microsoft significantly advanced AI image creation by integrating OpenAI's DALL-E AI image creator into Bing Chat, creating the Bing Image Creator. This tool generates images based on user's written text. In June 2023, Bing Image Creator was updated to include “Precise” and “Balanced” modes, providing users with more control over the image generation process.
In April, Microsoft took another leap making its Designer tool with OpenAI's DALL-E AI accessible to everyone. This innovative tool also can generate images from text descriptions, offering a unique way to create personalized designs.
Around the same time, NVIDIA was making strides in generative AI research, developing new techniques to enhance the quality and realism of AI-generated images. OpenAI further expanded the possibilities of AI in image creation with the release of ShapE, a generative model capable of creating 3D models from text descriptions.
Stability AI introduced StableStudio in May, an open-source version of its DreamStudio web app. StableStudio uses Stability AI's Stable Diffusion model to generate images from text prompts, enabling users to create a wide range of images. Stability AI further enhanced its Clipdrop tool by launching Reimagine XL, a feature that allows users to create multiple variants of an image using Stable Diffusion.
Ai-Based Image Editing
AI's capabilities are not limited to image creation but also extend to image editing, as demonstrated by a new unreleased tool called DragGAN. Developed by a team of researchers, DragGAN uses a point-based editing technique that allows users to make precise modifications to specific areas of an image without affecting the entire image.It offers users the ability to edit images as if manipulating a 3D model, resulting in more granular control and precision.
The system generates new pixels in response to user input, rather than merely altering existing ones, enabling users to deform an image with precise control over pixel relocation.The researchers behind DragGAN plan to extend this point-based editing technique to 3D generative models, promising an even wider array of applications.
Adobe also made a significant contribution to AI image creation and manipulation by integrating its AI image generator, Firefly, into Photoshop. Firefly allows users to extend images, add or remove objects, and enhance their creative workflow using text prompts.