15 Seconds of Speech is All It Takes to Clone a Voice with OpenAI’s Voice Engine

OpenAI’s Voice Engine: Cloning Voices in a Flash, But Can We Trust It?

OpenAI’s new Voice Engine can generate eerily realistic synthetic voices from a mere 15-second audio sample. This powerful tool, under development since late 2022, can not only mimic someone’s voice in their native language, but even translate it to create speech in entirely different languages!

While early access granted to select companies highlights the potential of Voice Engine in various fields, ethical concerns loom large. The ability to impersonate someone’s voice raises serious questions about potential misuse.

OpenAI is taking steps to mitigate these risks. They’ve established usage guidelines, requiring explicit consent for voice use and adding watermarks to track audio origin. However, enforcing these guidelines and preventing the technology’s potential cloning remain challenges.


Curious to hear how realistic these voice replications are? Check out audio samples demonstrating Voice Engine’s capabilities on The Verge’s website.

Show More

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Articles

Back to top button