Microsoft Bing’s AI Chatbot Can Now Generate Images with DALL-E

Bing AI

Microsoft has announced its new Bing AI chatbot will now feature the ability to generate AI images.

The Bing Image Creator is powered by OpenAI’s DALL-E and will allow users to create any image simply by describing what it is they want to see to the chatbot.

“For those in the Bing preview, Bing Image Creator will be fully integrated into the Bing chat experience, rolling out initially in Creative mode,” writes Yusuf Mehdi, Microsoft’s head of consumer marketing.

“By typing in a description of an image, providing additional context like location or activity, and choosing an art style, Image Creator will generate an image from your own imagination.

“It’s like your creative copilot. Just type something like ‘draw an image’ or ‘create an image’ as a prompt in chat to get creating a visual for a newsletter to friends or as inspiration for redecorating your living room.”

Image Creator

The Bing chatbot currently has three modes: creative, balanced, and precise. While precise mode gives more accurate answers, creative mode’s responses are more “original and imaginative” and it’s in creative mode where users can generate AI imagery.

Microsoft is also making Bing Image Creator available in its web browser Microsoft Edge which it claims is “the first and only browser with an integrated AI-powered image generator.” Edge will receive a new icon for the image generator in the sidebar.

Microsoft has said its image generator is using an “advanced version” of DALL-E with no further details given. Ars Technica notes that this could be a harbinger of a more powerful version of DALL-E that will be released soon.

Using AI Image Generators Responsibly

The dangers of AI image generators are obvious and well-documented. Microsoft has moved to allay some of those fears by slapping a modified Bing icon on the bottom left corner of each AI image to indicate that it is not a genuine picture.

Microsoft has also added its own safeguards on top of OpenAI’s protections to curb extreme use of its technology.

“For example, we have put controls in place that aim to limit the generation of harmful or unsafe images,” adds Mehdi.

“When our system detects that a potentially harmful image could be generated by a prompt, it blocks the prompt and warns the user.”