Microsoft has filed a patent application for AI that rewrites live speech within a few seconds, then redraws the speaker’s lips to match. Dexerto reported the filing (US 2026/0341309 A1), which was published on 29 October 2026.
The swapped words would come from preset rules or a viewer’s personal profile. That means two people watching the same stream could hear different versions of what the speaker said, and the published text includes no notice that an edit happened.
In the system the application describes, AI agents monitor a live stream, flag targeted words, and replace or remove them. A generative AI component then smooths the surrounding audio and adjusts the speaker’s mouth and facial movements so the picture matches. The filing defines “near real-time” as within a few seconds, giving 1 to 10 seconds as an example rather than a measured result. It ties that window to the delay broadcasters already use to filter profanity by hand.
Live translation heads the list of suggested uses, followed by filler words like “um” and “ah”, stutters, mispronunciations, regional slang such as “y’all” and “gonna”, and profanity. A separate claim covers translation: words in a language missing from your profile are swapped for translated ones, and the video is redrawn, including the lips. The benefits section also mentions “instant fact-checking”, with replacement words pulled from internet searches or databases.
According to the application, a profile could include age, languages, region, religion, ethnicity, education, and voice or video samples. That list describes what a profile could hold, not data any Microsoft product collects. One dependent claim keeps the original recording as “ground truth”.
Without clear disclosure, viewers could mistake an AI-modified presentation for an accurate record of what was originally said.
TechTimes
Those worries are about what the design could enable; no documented incidents exist. Microsoft hasn’t confirmed plans to ship anything and didn’t respond to Futurism’s request for comment. Its shipping Interpreter feature in Teams translates only audio and doesn’t touch the speaker’s face. Microsoft isn’t the only one patenting live AI edits. Sony’s patent for AI that can pause, blur, mute, or replace sensitive moments in games describes a similar idea with parent-set rules.
The application isn’t through yet. On 29 July 2026, the examiner rejected the claims as filed, a routine step usually followed by amendments, and Microsoft can respond.
