AI headphones filter out noise so you hear one voice in a crowd

A University of Washington team has developed an artificial intelligence system that lets a headphone user single out the voice of a specific person while suppressing all other sound in noisy places like restaurants or construction sites.
By looking at a person speaking for three to five seconds and pressing a button to record their speech, the system, called “Target Speech Hearing,” then cancels all other sounds in the environment and plays just the chosen speaker’s voice in real time even as the listener moves around in noisy places and no longer faces the speaker.