Meta Platforms has presented a new model of generative artificial intelligence, which is capable of generating speech.
This model is called Voicebox. In the press release of the technology giant, which was published on the occasion of the presentation of the configuration of generative artificial intelligence, it is indicated that with the help of a new AI-based product, users will be able to solve tasks such as audio editing, sampling, and stylization.
Voicebox can create high-quality audio clips and edit pre-recorded audio, for example, remove car horns or a dog barking from the final version of the audio recording. Also, in this case, the technology giant guarantees the preservation of the content of the audio material and its style. The new model of generative artificial intelligence is multilingual, as a result of which it can reproduce speech in six languages, including English, Polish, Portuguese, Spanish, French, and German.
A press release published by the technology giant contains information that Voicebox features include text-to-speech conversion using audio samples, the duration of which does not exceed two seconds. Also, this tool can recreate fragments of speech for the purpose of editing and noise reduction. Another function of Voicebox is to read text with a human voice in any of the six languages.
From the point of view of development prospects, this model of generative artificial intelligence has the potential to expand application options. For example, using the tool, virtual assistants can start talking in a voice that will not differ in any way from natural human speech. Also in the space of the metaverse, Voicebox for people who have vision problems can voice written messages from their friends.
Another potentially possible application of the new model of generative artificial intelligence is the creation and editing of audio tracks. In addition, Voicebox will help to help people who speak different languages communicate.
The press release of the technology giant indicates that the new tool is an important moment of advancement in the research of artificial intelligence capabilities. The company also announced its intention to continue to expand the scale of AI applications in the audio space.
Experts say that generative artificial intelligence is already a significant tool for a brand’s customer service, which is able to determine the emotional state, give advice and make transactions from the initial to the last stage of this process. In the US, industry research has shown that 61% of consumers are convinced that voice assistants will be able to replace people in terms of intellectual level and reliability.
Alphabet and Microsoft are also considering options for using generative AI for voice applications. In April, these companies announced the development and implementation of artificial intelligence tools that help in creating content, collaboration, and improving search results in terms of their compliance with individual needs within a specific query.
As we have reported earlier, Meta Says Generative AI and Metaverse to Evolve Together.