OpenAI has revealed the small-scale preview of their latest AI model, “Voice Engine”, which generates natural-sounding synthetic voices. It is a voice cloning AI model that OpenAI claimed they have been working on for a long time.
To clone a voice this Voice Engine needs only two things a sample of 15 seconds and a text as a script to follow.
OpenAI mentioned on their blog that “We first developed Voice Engine in late 2022, and have used it to power the preset voices available in the text-to-speech API, ChatGPT Voice and Read Aloud.” After completing various private testings OpenAI concluded that this Voice Engine can be used for a lot of beneficial purposes across many industries. One best example of its use is Age of Learning, which used this Voice Engine and GPT4 to educate students and create content for a wider audience. Moreover…
- It can help non-readers and children to learn from a natural voice
- It can help creators and businesses to reach a wider audience by translating their content into various languages.
- It can support people with speech issues to communicate using non-robotic and unique voices in different languages.
See Related: Google’s Latest AI Can Play Video Games With You While Following Your Commands
Safety And Risks of Voice Engine
Voice Engine has numerous benefits, yet like everything else, it also has negative effects. It can be used to make fraud calls and scam small businesses and families by demanding money from them using the voice of someone they know. It can create the voice of professional voice artist in a way that can harm their reputation and jeopardize their income.
To avoid all these and other voice-related scams and misuse, OpenAI is taking a cautious approach before the broader release of Voice Engine. They also have some policies and rules to use their Voice Engine which need to be followed. OpenAI shared some steps to be taken by the authorities to avoid the risks related to AI on their blog.
