Microsoft is expanding its artificial intelligence (AI) offering to the public with its own variant of the Dall-E image generator. It comes from Open AI and generates images from text input, similar to Midjourney and Stable Diffusion. Microsoft has access to a further developed version of Dall-E and integrates this offer under the name Bing Image Creator Preview both in the preview of its search engine chat and in the Internet browser Edge.
Access is gradually being released to Microsoft accounts around the world, but will only work in English for the time being. Microsoft has this on announced Tuesday. The data company emphasizes that it has installed additional precautions to prevent certain images. Users entering such commands should be warned. For example, Midjourney is known to have words like “placenta”, “fallopian tube”, “mammary glands”, “sperm”, “womb”, “urethra”, “cervix”, “hymen” or “vulva” (each in English). language) prohibited.
Flood of images in search results
Access is either via the sidebar of the Edge browser or via the preview version of the Bing chat, initially only in the “creative” variant. This variant takes more liberties, but also makes significantly more mistakes than the alternatives “balanced” and “precise”. The use of the Bing Image Creator Preview is free of charge until further notice, but a Bing logo is emblazoned in one corner of each generated image.
In addition, the search engine now offers so-called Stories and Version 2.0 of the Knowledge Cards. Both rely heavily on images and videos. Stories are search results that are intended to introduce the respective topic with videos and images. Microsoft describes the Knowledge Cards as an “AI-powered infographic-inspired experience that offers fun facts and essential information at a glance”. The new features of version 2.0 include “interactive, dynamic content such as tables, curves and time series”.
(ds)
