Microsoft has filed a patent application for an AI system that could change spoken words during live broadcasts. The technology is designed to identify speech automatically, then remove or replace words while adjusting the speaker’s mouth movements to match the edited audio, according to Futurism.
The patent describes possible uses including translating broadcasts into other languages, removing filler words and correcting verbal slip-ups. Broadcasters or viewers could also customize the system to filter profanity, offensive language or selected phrases.
Microsoft says the technology could be used for lectures, award ceremonies, sporting events and live television programs. The original recordings could also be retained as a record of what was actually said.
The filing describes a proposed technology; it does not establish whether or when Microsoft will release it.
Sources: Computerworld.