Microsoft has filed a patent application for an artificial intelligence-based system capable of editing live video streams and augmented reality in real time. The new technology makes it possible to correct or completely replace a speaker's speech while simultaneously adjusting facial expressions and lip movements to create a natural visual and audio effect.
The system's operation is ensured by special AI agents that process video and audio signals with minimal latency, creating smooth transitions before and after modified fragments. Modifications are made based on predefined parameters or viewers' personal profiles.
According to the patent documentation, the technology can be used during broadcasts of lectures, presentations, sports events, and TV shows. Among the main use cases indicated are:
The introduction of such a tool allows viewers to customize someone else's speech to their own preferences, but this raises concerns about the further isolation of users in personalized "information bubbles" and the distortion of reality during live broadcasts.
The patent states that to prevent abuse and distortion of facts, the original audio and video recording of the event will be stored as "ground truth." Currently, Microsoft has not provided official comments regarding the timeline for the practical implementation of the development.
Нагадаємо, No Second Chances: Worlds 2026 Playoffs Preview and Swiss Stage Results.
