OpenAI creates an AI that clones your voice in just 15 seconds (but keeps it under lock and key because of the risks inv

[tasnimsanika00]

OpenAI has unveiled a new AI tool, Voice Engine, which claims to be able to clone anyone's voice in an incredibly realistic way from an audio sample of just 15 seconds in length. However, despite how amazing the tool is, ChatGPT's parent company has preferred to keep it under lock and key and not make it available to the general public due to the enormous risks it entails. , especially when we are on the threshold of the absolutely decisive elections all over the world.

Voice Engine actually began development in late 2022, and an early version of the tool was actually used in ChatGPT , OpenAI’s flagship app. However, this is a tool that has never been made available to the general public , largely due to the “cautious and informed approach” that the company led by Sam Altman is taking when undertaking a broader launch of the app.

"We hope to start a dialogue about the responsible development of synthetic voices and how society can adapt to these new advances ," OpenAI explains in a post on its corporate blog . . “Based on these conversations and the results of small-scale tests, we will make more informed decisions about how to deploy this technology on a larger scale,” the American startup adds.

In the post published on its corporate blog, OpenAI shares several real-life examples of use from partners who have already had access to Voice Engine. and have been able to integrate this technology into their applications and products.

Education technology company Age of Learning , for example, uses Voice Engine to generate scripted voiceovers, while HeyGen , a storytelling-focused app, uses OpenAI software to generate translations of previously recorded content in a way that preserves the original speaker’s accent and voice.

OpenAI has chosen not to launch the Voice Engine tool in a mass release
Meanwhile, a research team at the Norman Prince Neurosciences Institute in Rhode Island used a 15-second audio clip from a young girl (which was remarkably low-quality) to restore the protagonist's voice, which she had lost due to a brain tumor.

“We have chosen not to launch this technology at this time in order to strengthen society’s resilience against the challenges posed by generative artificial intelligence models whose results are increasingly convincing,” OpenAI argues. “And we encourage steps such as phasing out voice-based authentication methods for accessing bank accounts and other so bulgaria whatsapp list urces containing potentially sensitive information,” the company asserts.

OpenAI also calls for the exploration of “policies specifically aimed at protecting the use of individual voices in AI” and for “educating the general public to properly understand the capabilities and limitations of AI, including the potential for misleading content to be generated with this technology.”

The content generated by Voice Engine is provided with watermarks in all cases. , allowing the origin of any artificially generated audio to be traced, OpenAI emphasizes. “Our terms of use with partners who already have access to Voice Engine require explicit and informed consent from the original speaker, and we do not allow developers we collaborate with to rely on individual user voices to artificially create voices,” OpenAI notes.

While OpenAI has chosen not to make Voice Engine available to everyone, voice generation tools from companies like ElevenLabs are, however, freely accessible to everyone . However, ElevenLabs has put in place safeguards to prevent the creation of voices that imitate candidates actively involved in electoral processes.

Newsletter
Subscribe to our newsletter!