First, OpenAI offered a tool that allowed people to create digital images by simply describing what they desired to see. He then built similar technology that generated full-motion video like something out of a Hollywood movie.
Now he has unveiled technology that can recreate someone’s voice.
On Friday, the high-profile artificial intelligence startup said a small group of corporations is testing a brand new OpenAI system called Voice Engine that can recreate an individual’s voice from a 15-second recording. If you upload a recording of yourself and a paragraph of text, it could actually read the text in an artificial voice that sounds just like yours.
The text doesn’t should be in your native language. For example, for those who speak English, it could actually reproduce your voice in Spanish, French, Chinese and lots of other languages.
OpenAI is just not making the technology available more broadly since it remains to be trying to grasp the potential threats. Like image and video generators, a voice generator might help spread misinformation on social media. It also can allow criminals to impersonate other people online or over the phone.
The company expressed particular concern that one of these technology may very well be used to compromise voice security that controls access to online banking accounts and other personal applications.
“It’s a delicate matter and it’s important to get it right,” OpenAI product manager Jeff Harris said in an interview.
The company is exploring ways to watermark synthetic voices or add controls to forestall people from using the technology with the voices of politicians or other distinguished figures.
Last month, OpenAI took an analogous approach when unveiling its video generator, Sora. He showed off the technology but didn’t make it public.
OpenAI is one in all many corporations that have developed a brand new form of artificial intelligence technology that can quickly and simply generate synthetic voices. These include tech giants like Google in addition to startups like New York-based ElevenLabs. (The New York Times sued OpenAI and its partner Microsoft over copyright infringement claims regarding text-generating artificial intelligence systems.)
Companies can use these technologies to generate audiobooks, provide voice to online chatbots, and even create an automated DJ radio station. Since last yr, OpenAI has been using its technology to support a talking version of ChatGPT. It has long offered enterprises a spread of voices that might be utilized in similar applications. They were all built from clips provided by the voice actors.
However, the company has not yet offered a public tool that would allow individuals and firms to recreate voices from a brief clip, as Voice Engine does. The ability to breed any voice in this manner, Harris said, makes the technology dangerous. In his opinion, this technology could also be particularly dangerous in an election yr.
In January, New Hampshire residents received automated messages advising them to not vote in the state’s primary election with a voice that was likely artificially generated to sound like President Biden’s. The Federal Communications Commission later banned such calls.
Harris said OpenAI has no immediate plans to generate income from the technology. He said the tool may very well be especially useful for individuals who have lost their voice as a consequence of illness or accident.
He showed how this technology was used to recreate a lady’s voice after it was damaged by a brain tumor. He said she will be able to now speak after a brief recording of a presentation she once gave when she was a highschool student.
