Microsoft's new AI creates live images while you talk
We all know that Microsoft is betting big on AI, we've already seen it with Copilot in almost every software and in their operating system, and now they want to tackle image generation.
However, Copilot itself, based on Microsoft Designer, can create many images, but thanks to the discovered patent, it seems that Microsoft wants to go much further in this regard.
According to the patent registered with the United States Patent and Trademark Office, this document talks about an AI-powered system for converting live audio into images.
That is, it is a system capable of generating images in real time, while, for example, a meeting is taking place in Microsoft Teams or similar programs.
During the meeting, when users speak, these words will be captured by a microphone and then converted into text. The text will then be broken down into sentences and each part will be summarized using a language model to generate prompts for generating images.
This way, depending on the topic being discussed, users will also have images created using these instructions so that they can follow the meeting in a better way.
“When images are used to complement verbal communication, they can help clarify concepts and make them easier to understand, which can be especially helpful for people who learn best through visual means,” Microsoft says in describing the idea behind the technology.
It would be a perfect feature for Microsoft Teams, as these AI-generated images would be displayed on the live screen while the audio is still on.