OpenAI presented the voice copying AI Voice Engine – Kommersant

American company OpenAI announced a new Voice Engine neural network, which can copy a 15-second voice excerpt and voice the written text, trying to most accurately convey the speech characteristics and voice of the speaker. The company claims that it will prohibit the use of a neural network to impersonate another person.

OpenAI believes that the new AI will be useful for translating content into other languages ​​in the voices of the authors. The neural network will also be able to help children and people with reading problems using natural and emotional voices. Voice Engine can be effective for therapeutic purposes in supporting users with speech impairments, as well as in voice restoration for speech problems.

Previously OpenAI announced, which will make the realistic AI video generator Sora available to the general public within 2024. Also Bloomberg reported that OpenAI is going to attract Hollywood film studios to use the Sora neural network.

Evgeniy Fedunenko