Microsoft has filed a patent for an artificial intelligence technology capable of modifying a person’s words in a video in nearly real time while synchronizing their facial expressions and lip movements accordingly.
The system would utilize AI agents to locate specific words or phrases, delete or replace them, and then edit the video and audio to ensure a seamless transition.
Among the potential scenarios outlined by Microsoft is the translation of live broadcasts into other languages. In such cases, viewers would not only hear the translation but also see the speaker’s lip movements adjust to match.
Additionally, the technology could remove filler words, stuttering, mispronounced terms, or offensive language. The system would be able to select replacements based on established parameters or a specific user profile.
The patent also describes the capability to alter colloquial expressions and other phrases that a viewer might prefer not to hear while watching a broadcast.
Note: This article is based on publicly available sources and reflects information available at the time of publication. Details may change; check the original or official source for updates.