Meta, formerly the Facebook company, recently announced creation of a new AI model to enhance speech generation.
Voicebox is claimed to be “the most versatile AI speech for generations,” and intended to be used to help speech editing, noise reduction, text-to-speech synthesis, and cross-lingual style transfer. Meta believes Voicebox could allow people who are blind or with low vision “to hear written messages from friends in their voices.”
Forbes questions how an AI tool such as Voicebox will respond to audio samples that contain atypical speech patterns, such as stutters. Voicebox’s accessibility is still being evaluated.
Meta is currently withholding Voicebox’s release to the public due to current ethical discussions regarding AI-based technologies.
For more information, please read Forbes article on Meta’s Voicebox AI model.