From Neowin Media Group: 2024-03-31 01:26:00

OpenAI has developed a powerful AI voice cloning model called Voice Engine that can create realistic voice expressions with just a 15-second audio sample. The company is concerned about the potential for misuse and is cautious about releasing the technology to the public. OpenAI is conducting tests with trusted partners to assess the benefits and risks of wider deployment.

Voice Engine’s capabilities have impressed users and are being tested for various applications such as text-to-speech API and ChatGPT Voice. OpenAI is exploring the responsible deployment of synthetic voices and considering policies to prevent misuse, such as verifying original speakers and creating a “no-go list” for voice re-creation of prominent figures.

The company emphasizes public dialogue on AI technologies and suggests phasing out voice-based authentication for sensitive information, protecting individuals’ voices in AI, and educating the public on AI capabilities and limitations. OpenAI’s Voice Engine shows promise in advancing voice cloning quality compared to existing tools like ElevenLabs.



Read more at Neowin Media Group: OpenAI can clone your voice with little to no input, and it wants to talk about it with you