Microsoft Patents AI System to Edit Speech and Faces in Live Video
Reported by Futurism1 source ↓
Microsoft filed a patent describing an AI system that could alter words in live video and adjust a speaker’s face and mouth to match. The patent lists translation and speech filtering as possible uses.
Key points
- The system could remove or replace words and edit a speaker’s mouth and face to match
- The patent lists translation, filler-word removal and filtering offensive or misused words as possible uses
- The patent does not establish that Microsoft will release the system
- Microsoft did not immediately respond to Futurism’s request for comment
Sources
Details
Futurism reports that Microsoft filed a patent for AI agents that would modify video and audio just before and after selected words to make edits sound and look continuous. Changes could be based on preset parameters or user profiles, and replacement words might come from an internet search.
The patent describes uses including real-time translation, removing filler words or stutters, and filtering terms a viewer considers inappropriate. It also says original audio and video may be stored as “ground truth.” The patent covers possible applications in lectures, television productions and live sports; it does not confirm that Microsoft plans to build or release the system.
Written from coverage as of
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh