[ad_1]
Audio deepfakes, or AI-generated audio that impersonates somebody, aren’t new — they usually’re the rationale that OpenAI, the corporate that introduced AI chatbots to the mainstream, is holding again on releasing its newest providing.
OpenAI introduced on Friday that it was selecting to preview, however not broadly launch, text-to-speech sensible voice generator Voice Engine due to “the potential for artificial voice misuse.”
In a blog post, the corporate outlined Voice Engine’s capability to take a 15-second pattern of somebody’s voice and emotionally and realistically mimic it as directed by a textual content enter.
“If in case you have the precise audio setup, it is mainly a human-caliber voice,” Jeff Harris, a product lead at OpenAI, told Bloomberg. “It is a fairly spectacular technical high quality.”
Associated: Deepfake Scams Are Becoming So Sophisticated, They Could Start Impersonating Your Boss And Coworkers
OpenAI has been privately testing Voice Engine since growing it over a 12 months in the past and has recognized that it may be “used for good” in its weblog publish.
One software that the corporate previewed, Voice Engine helps people who find themselves non-verbal by giving distinct voices throughout many languages. Livox, an alternate communication app, has already began utilizing Voice Engine for that function, based on OpenAI.
Voice Engine may additionally translate movies and podcasts into different languages and supply studying help to kids and non-readers with audio content material.
Associated: Tennessee Just Passed a New Law to Protect Musicians From a Growing AI Threat
OpenAI pointed to its AI safety and voluntary commitments insurance policies when stating its rationale for previewing, and never releasing, Voice Engine to the general public. The preview was meant to showcase Voice Engine’s capabilities whereas additionally emphasizing “the necessity to bolster societal resilience towards the challenges introduced by ever extra convincing generative fashions,” the corporate said.
Artificial voices have captured curiosity within the startup world, with AI voice cloning firm ElevenLabs valued at $1.1 billion in January after launching in beta solely a few 12 months in the past. The know-how has additionally come beneath hearth for the brand new energy it provides cybercriminals, who can use it to impersonate people or entry funds by way of voice-based banking.
OpenAI previewed an AI video generator known as Sora final month that creates sensible movies from prompts.
Associated: ‘This Is a Serious Problem’: Mr. Beast Slams AI Deepfake
[ad_2]
Source link
