Microsoft has announced the release of the DALL-E 3 text-to-image conversion model for all Bing users, available now in Bing Chat and Bing Image Creator.
DALL-E 3 is an AI-powered system that generates images from textual prompts, with Microsoft describing the results as both realistic and creative.
The company has highlighted that the new system brings improvements to overall image quality and detail, including added precision for human hands, faces, and text.
Last week, Microsoft introduced the latest version of its AI-powered image generator for Bing Enterprise users, and it is now accessible to all users.
Bing gains access to DALL-E 3 even before ChatGPT, and this month, paying ChatGPT users will also gain access to DALL-E 3. Microsoft’s search engine is expected to serve as a popular image creation tool for some time, as people leverage this technology for illustrations, social media content, backgrounds, and more.
DALL-E 3 is the third iteration of OpenAI’s image creation model, designed to be user-friendly, it integrates with Bing Chat and ChatGPT, allowing you to create and refine images through conversation with the chatbot rather than attempting to improve your prompt.
OpenAI has introduced new safety measures in DALL-E 3. The model does not recreate images of public figures, and it refrains from generating images promoting hate speech or illegal activities.
Microsoft adds watermarks to all AI-generated images, identifying them as AI-generated, and all images generated by Bing Image Creator include an invisible digital watermark compatible with C2PA specifications, providing information about the image’s creation time, date, and source verification.
Microsoft has also implemented a content moderation system adhering to Bing’s terms of service and community guidelines, excluding images containing nudity, violence, hate speech, or unlawful activities.
Microsoft plans to use the DALL-E technology beyond Bing, developing a Paint Cocreator tool for AI-powered image creation in the Paint application for Windows.
Regardless of text-based prompts, the model also offers suggestions for appropriate modifications and additions to the image. Microsoft encourages users to provide detailed text prompts for accurate results from DALL-E 3.
