Patent documents show that Microsoft is working on a new AI feature that creates images based on live audio during meetings to make communication visual.
A picture says more than a thousand words: Microsoft seems to be taking this saying quite literally. Microsoft has filed a patent for a new AI technology that can convert live audio to images. Among other things, the technology could serve to provide Teams meetings with images.
The technical patent document explains how the system works. The AI listens in during meetings and converts the audio to text for itself. Based on those text summaries, the model then generates images associated with the conversation. This is all done in real-time so that the images follow the content of the meeting as closely as possible.
Seeing, hearing and speaking
Filing a patent does not automatically mean Microsoft plans to launch the technology for the general public. But the software giant does see it as a useful addition for video calls.
“When images are used to complement verbal communication, they can help clarify concepts and make them more understandable, which can be especially beneficial for people who learn better with visual aids,” Microsoft explained its rationale behind the technology.
New updates have already given Microsoft Copilot eyes and the ability to speak. Microsoft is looking for ways to make its AI assistant relevant to everyone. It is succeeding with varying degrees of success: the number of Copilot users is increasing month by month, but not everyone is yet convinced of its usefulness for their work, including Salesforce CEO Marc Benioff.
read also