Voicebox is Meta’s new AI voice generator, which they claim is the most advanced and powerful tool of its kind.
In a blog post released on Monday, Meta declared Voicebox as a breakthrough in generative AI for speech, far exceeding what it was ‘specifically trained to accomplish’. But the tech giant is holding off on releasing it to the public for fear of ‘potential risks of misuse.’
While programs like OpenAI’s ChatGPT and Google’s Bard can generate text from a prompt, and AI image generating programs like Midjourney and Stable Diffusion can create images from a text prompt, instead of creating a picture or a passage of text, Voicebox is able to generate audio files from text prompts, and so much more.
Apart from creating audio files from scratch, it can also modify the audio files it creates, as well as audio samples input into the program in a variety of ways.
Voicebox is able to convert text or audio files into English, French, Spanish, German, Polish, and Portuguese. Imagine being able to tear down language barriers and having almost real time conversations in a different language with someone you’d have never been able to communicate with before.
Not only is it able to translate text and audio files, but you can also give it a sample of your voice and it will translate your input into another language, using your voice!
It can also remove background noise from audio files, as well as allow you to seamlessly edit voice clips without having to rerecord the whole thing.
The applications of such an advanced audio generation program are so numerous and far reaching in terms of ushering in a more inclusive, efficient and exciting new era.
But on the same token, the negative implications of such a program are is equally as frightening. Now just imagine a program that would allow anyone to take a sample of someone’s voice and clone it and use their voice to say whatever they want them to without their permission, to spread misinformation, and to use for nefarious purposes .
Right now, we’re living in the wild west of AI where anything goes, and the technology is advancing at a far faster rate than we are able to regulate it.
As fantastic as this program sounds, it’s probably just as well that Meta is holding off on releasing it to the public before it can figure out how to implement safeguards to try and ensure the use Voicebox’s powers for are good instead of evil.