Microsoft continues its race for artificial intelligence by integrating image-generating AI into Bing and Edge. Called Bing Image Creator, this module is capable of creating an illustration from scratch from a textual description.
Microsoft continues its offensive in the field of artificial intelligence! After integrating the GPT-4 language model in its Bing search engine and in its Edge browser, then announcing the launch of a Copilot assistant for its Microsoft 365 suite (see our article), the Redmond firm is taking a new step with Bing Image Creator (“Bing image creator”), an image generator directly integrated into its chatbot. Thanks to it, the user can, from a short textual description, create corresponding visuals. And the result is quite impressive! A small revolution that will probably boost the popularity of the search engine since, since the launch of its chatbot, Microsoft claims more than 100 million conversations. With Bing Image Creator, the counters may panic!
Bing Image Creator: an image thanks to a description
Bing Image Creator works like DALL-E and Midjourney: all you have to do is describe the image you want to obtain for the AI to generate four proposals. If one of them interests the user, he only has to click on it to obtain a clearer and more precise rendering. Otherwise, he can modify his request by trying to be clearer. “By entering an image description, providing additional context such as location or activity, or choosing an art style, Image Creator will generate an image from your own imagination. explains Microsoft in his press release.
At this time, Bing Image Creator is only available to users registered for the Preview version of Bing Chat, in the mobile and desktop versions of the search engine and the Edge browser. For the latter, just click on the tool icon located in the right sidebar or ask directly in the chat. Otherwise, it is also possible to test the tool at this address, by logging into their Microsoft account. The Redmond firm is also quite proud to indicate that Edge is the first browser equipped with an image-generating AI. “Historically, search was limited to images that already existed on the web. Today, there are virtually no limits to what you can search and create”she triumphs.
Bing Image Creator: new features galore
Bing Image Creator being available in a version preview, there may be some hiccups. The AI is currently only accessible in the “Creative” mode of Bing Chat – to better adapt the chatbot to their needs, users can configure the chatbot according to three different types of personalities (see our article) –, but Microsoft intends to extend it to “Balanced” and “Precise” modes later. Likewise, it only includes requests in English, but additional languages, including French, will be added later.
This isn’t Microsoft’s only addition to the chatbot. Indeed, the company announces a “Stories” function, which offers “a more engaging way to search and interact with content, featuring images and short videos.“Similarly, the “Knowledge Maps” feature is moving to version 2.0. Inspired by infographics and powered by AI, it “provides fun facts and key information at a glance”. The update makes it possible to offer “interactive and dynamic content”like charts, tables, timelines and more.
Bing Image Creator: a tool not to be put in all hands
Microsoft’s goal is to develop a multimodal tool, that is to say having the ability to manipulate text, images and other forms of content just as well to generate others of various kinds. Note that Bing Image Creator is much more accessible than Midjourney, where you have to enter prompts in public Discord rooms – and find your images among the countless requests made at the same time.
Of course, it is necessary to set limits and develop safeguards with this type of AI, in order to prevent it from being used to generate inappropriate or illicit content – violent, sexual, child pornography, or even deepfakes. of public figures. A tricky task, considering how the Conversation version of Bing had gone off the rails. Also, Microsoft explains that when the user’s request is likely to result in a problematic image, a warning is sent to him and the command is blocked. The company claims to have worked in partnership with OpenAI – at the origin of ChatGPT and DALL-E – and to have added additional filters itself. Additionally, images generated by Bing Image Creator will be automatically tagged with a Bing icon to let users know the artwork is AI-powered.