OpenAI’s voice synthesizer can clone your voice in 15 seconds of audio

The technology, known as Voice Engine, is currently undergoing a limited preview phase, with applications such as educational aids, language translation for podcasts and assisting non-verbal individuals

OpenAI has made significant progress with ChatGPT conversational AI and Sora AI video creator in the past year.

The latest addition to its arsenal is Voice Generation, a tool capable of crafting synthetic voices from just 15 seconds of audio.  

This technology, known as Voice Engine, is currently undergoing a limited preview phase, with applications such as educational aids, language translation for podcasts and assisting non-verbal individuals.

While the generated samples exhibit impressive quality, concerns about potential misuse, such as spreading misinformation or unauthorized voice replication, have prompted OpenAI to tread cautiously.  

The organization aims to foster discussions on the ethical use of synthetic voices and plans to make informed decisions based on feedback and research outcomes on widespread deployment.

Given the upcoming major elections and the rapid advancement of generative AI tools, the issue of trustworthiness in various AI-generated content forms – audio, text and video – remains a pressing challenge. 

Source: Newsroom