Bing Chat was already leading the AI chatbot space with OpenAI’s latest GPT-4 technology and full access to the web. On Tuesday, Microsoft announced its Bing Image Creator, a text-to-image generator powered by an advanced version of OpenAI’s DALL-E model.
The image creator will exist within Bing Chat, meaning that within the same chat session, users will be able to generate text and image responses. This is a big contrast from ChatGPT and DALL-E 2 which live on two different websites.
The experience will first be limited to the Creative mode in the chat. To generate an image, all a user will have to do is input what they would like to see rendered with as much detail as possible. Users can include specific details such as location, activity and artistic style.
Bing’s Image Creator will also be accessible within Microsoft Edge’s sidebar, the same way Bing Chat is now. The image generator will have its own icon on the sidebar, where users can quickly type in their image prompt to get an image rendered, regardless of where they are browsing on the web.
To promote ethical image generation, Bing will automatically block prompts that can be harmful and unsafe. Bing will also include a watermark in the corner to clearly delineate that the image has been AI generated to avoid confusing the public.
Bing Image Creator will be rolled out to Bing preview users starting today. If you haven’t gotten access to Bing Chat yet, you can join the waitlist.
Even if you have no interest in experiencing Bing Chat, you can still preview the image-to-text generator by visiting Bing’s Image Creator preview.
Microsoft is also releasing new AI-powered visual stories and knowledge cards to all Bing users. When users search something on Bing, they will be met with quick, concise, interactive infographics including dynamic charts, graph, timelines and visuals regarding their search, according to Microsoft.