In some movies, for example Mission Impossible, we can see them showing the ability to clone a voice by only having to read a simple text. The matter is now seen to be a reality, when OpenAI today announced their synthetic voice generator, named Voice Engine.
Through the Voice Engine, it can generate a synthetic voice using just a 15-second voice or audio sample – and then users can use text-based input to generate audio using that voice.
OpenAI says the Voice Engine can help with a number of things, including enabling patients with speech impairments, in translations that preserve the same voice type, and a number of other things.
However, OpenAI also states that this Voice Engine offering can lead to some misuse problems – especially involving politics, where it can be misused to generate person-based voices for nefarious purposes.
For now, OpenAI offers access to this Voice Engine for some parties only. It is not yet known when OpenAI will open its access publicly to all.