It’s hard to get away from it now: OpenAI is everywhere. And to continue its integration into the Microsoft sphere, its DALL-E image-generating artificial intelligence will also be integrated into the Bing search engine, as noted in a company blog post.
Since a few days ago, Bing users can already use the services of ChatGPT (another famous OpenAI tool) and especially its new version GPT-4. Microsoft also claims that more than 100 million exchanges have already been made on the search engine. Therefore, the company decided not to stop there.
DALL-E is now available to “take the chatbot experience to another level by making Bing more visual,” in the words of Yusuf Mehdi, vice president in charge of Microsoft’s artificial intelligence project. He will note that the tool is not explicitly called DALL-E but “Bing Image Creator”.
Tool not yet developed
Eventually, the tool should be integrated into the same place as the conversational chatbot, as the images revealed by Microsoft show. But at the moment, the AI that generates images is only available as a preview on an external site. As Tech&Co noted, if we ask the Bing chatbot to generate an image for us, it responds “Sorry, but I can’t create images for you. I can only communicate via text.”
In preview, however, it is possible to test the Bing version of DALL-E. And the least we can say is that it’s still far from the realism of Midjourney. To test the tool, Tech&Co, for example, asked for a “realistic photo of the Paris protests with Emily in Paris drinking coffee.” If the Haussmannian fire and buildings are quite convincing, the characters disappoint much more: deformed faces, unreal hands or even eyes without irises.
Bing Image Creator will also integrate with Microsoft Edge, making the browser “the first to integrate image-generating AI,” according to the company. Through a button located on the toolbar, users will be able to obtain Bing Image Creator at any time during their browsing.
Source: BFM TV
