Meta unveils Voicebox AI to replicate the voices of your friends and loved ones

Meta has announced a new AI model called Voicebox, which can imitate voices. The model is capable of speaking in six languages. Meta described this development as a major advancement in AI technology for speech.

Although Meta shared this news, they have decided not to make the Voicebox model or its code available to the public due to the potential risks involved. This development could add to one of the artificial intelligence tools humans interact with daily.

In a research post, the company stated: "Generative speech models have many exciting applications, but because of the potential risks, we are not making the Voicebox model or code publicly available at this time. While we believe in sharing our research to advance AI, we also recognise the importance of being responsible and striking the right balance".

According to Meta, the model could be used to create custom software solutions, especially for visually impaired individuals who could leverage the service to hear messages in familiar voices or allow users to speak in foreign languages using their own voice. The technology also presents opportunities for creators to enhance audio tracks for videos or develop more natural-sounding voices for virtual assistants.

The company expressed its anticipation of seeing how other researchers will build upon their work. The service will be compatible with both Android and iOS applications. Possibly, Meta wants to embed enough security in the software before releasing it.

Meta has been emphasising its plans for AI development, particularly in integrating AI into its products. This is noteworthy since the company has been less visible in terms of consumer-facing AI products.