- Microsoft builds an image creator tool for its Bing Chat AI.
- The new AI feature uses the DALL-E model from OpenAI.
- The Bing Image Creator will be available on the web, mobile, and on Edge.
Microsoft is integrating a new feature into its Bing Chat to create images using AI. The new image creator uses the DALL-E model from OpenAI, which allows you to create virtually any image by describing to the chatbot what you want using text.
The Bing Image Creator will be available through the Chat experience. However, initially, the feature will only be accessible using the “Creative” style mode. In the future, the company plans to make the feature available in the “Balanced” and “Precise” modes.
As of right now, Bing with ChatGPT integration has three response tones (or modes), including “Creative,” “Balanced,” and “Precise.”
The “Balanced” mode allows the Bing AI chatbot to respond more neutrally, meaning it will try not to take sides on a specific topic. The “Creative” mode generates more playful and original responses. However, changing the tone (style) to “creative” may cause the chatbot to produce less accurate answers. And the “Precise” mode is the option that generates the most accurate response with more facts.
According to Yusuf Mehdi at Microsoft, you only need to describe the image, provide additional contexts like location and activity, and choose the art style for Bing to use the DALL-E model to create the image.
The feature is not only coming to the web but also to the Microsoft Edge browser. In the sidebar, you will notice a new icon to access the Bing Image Creator to generate images with AI.
The Bing Image Creator can be accessed from the Chat page by requesting the chatbot create an image. However, if you’re not in the new Bing experience, you can access the image creator from bing.com/create. You can also get early access to Bing chatbot with these instructions.
The text-to-image feature using AI is only available in the English language, but Microsoft plans to expand it to more languages in the future.
Although this new feature can open the experience of generating bad images, the company says that it has implemented safeguards into the Image Creator to prevent the misuse of the AI feature.
The feature is rolling now, but it might take some time to become available everywhere.