January 11, 2023 reportMicrosoft's VALL-E can faithfully reproduce a voice after listening to a three second recordingA team of researchers at Microsoft has demonstrated a new AI system that is capable of mimicking a person's voice after training with a recording just three seconds long.Subsequent work showed that it is capable of far more—not only can it mimic a voice, it can also simulate tone and even the acoustics of the environment in which the original recording was made.The app was taught to "listen" to a string of words and then to replicate its sound using Meta's Libri-light dataset, which has over 60,000 hours of recordings made by 7,000 people speaking in English.But because the app is still in its early stages, it is likely its functionality will improve over time.Microsoft has not made the source code for VALL-E public and likely will not do so, noting that it could be used in less than responsible ways—hoax recordings of politicians, for example."