Microsoft has filed a patent for an artificial intelligence system that could edit what people say during live video broadcasts, changing both their spoken words and lip movements to make the alterations appear natural.

The technology is designed to make changes within approximately 1 to 10 seconds while a video is streaming. It could replace selected words in a speaker’s audio and recreate their mouth movements to match the edited speech.

The patent application, titled “Near Real-Time AI Enhancement of Video or AR Presentations of Live Events” (US application 20260341309), describes a system that can modify speech before viewers receive the video.

What Microsoft’s AI Could Change

The proposed technology could remove filler words such as “um,” replace slang, correct mispronounced or unclear words, and filter out offensive language.

It could also personalize speech according to individual viewers’ profiles, including their age, location, and preferred language.

Another feature described in the filing involves translating speech into another language while synchronizing the speaker’s lip movements with the translated audio.

Concerns Over Manipulated Videos

The technology could create concerns about misinformation and the authenticity of live video content.

Since both audio and video would be modified before reaching viewers, audiences might have no obvious way to distinguish the original speech from AI-generated replacements.

The patent also reportedly includes only one dependent claim addressing the preservation of the original recording, leaving questions about how viewers could verify what a person actually said.

Is Microsoft Launching This Technology?

The filing is a patent application and does not confirm that Microsoft is developing a commercial product or plans to release the technology.

The company has continued expanding its AI offerings, including major upgrades to Copilot and the development of AI tools for medical diagnosis.