Meta Unveils Voicebox, a Generative AI Model for Speech Generation Meta, the parent company of Facebook and Instagram, has unveiled a new AI model called Voicebox that can induce realistic speech. Voicebox is trained on a massive dataset of audio recordings, and it can be used to produce synthetic voices that sound nearly mortal. Meta says that Voicebox could be used for a variety of purposes, similar as creating audiobooks or generating voice commands for virtual sidekicks. Voicebox is a generative AI model, which means that it can produce new affair from scrape. This is in discrepancy to discriminational AI models, which can only classify or prognosticate being data. Voicebox is trained on a massive dataset of audio recordings, which includes a variety of different voices, accentuations, and speaking styles. This allows Voicebox to induce synthetic voices that sound veritably n...