.elementor-panel-state-loading{ display: none; }



OpenAI Unveils Audio Feature That Read Texts, Clones Human Voices

OpenAI told reporters earlier this month that it decided against widespread deployment of this feature.

OpenAI is sharing early results from testing its ability to read words out loud in a convincing human voice. This highlights new frontiers for artificial intelligence and raises concerns about the risk of deepfakes.

The company is sharing early demos and use cases from a small preview of its text-to-speech model, called Voice Engine, with about 10 developers so far, a spokesperson said. Ta.

OpenAI told reporters earlier this month that it decided against widespread deployment of this feature.

An OpenAI spokesperson said the decision to scale back the release was made after receiving feedback from stakeholders, including policymakers, industry experts, educators, and creators. According to a previous press conference, the company originally planned to release the tool to up to 100 developers through an application process.

“We recognize that there are significant risks in producing audio that resembles people’s voices, and that is a top priority, especially in an election year,” the company said in a blog post Friday. Ta. “We work with U.S. and international partners in government, media, entertainment, education, and civil society to ensure we incorporate their feedback as we develop.”

Other AI technologies are already being used to disguise voices in some situations. In January, he encouraged people in New Hampshire not to vote in the primary in a fake but realistic phone call purporting to be from President Joe Biden. The incident stoked fears about AI ahead of a crucial global election.

Unlike OpenAI’s previous efforts to generate audio content, Voice Engine can create voices that sound like individual people, with specific rhythms and intonations. All the software needs is a 15 second audio recording of the person speaking to recreate the human voice.

During a demonstration of the tool, Bloomberg heard a clip of OpenAI CEO Sam Altman briefly explaining the technology in a voice indistinguishable from real speech, but entirely AI-generated. Met.

“With the right audio settings, your voice is essentially as good as a human voice,” said Jeff Harris, head of product at OpenAI. “It’s a very impressive technical quality.” But Harris said, “There are obviously a lot of safety considerations in terms of its ability to mimic human speech really accurately.”

One of OpenAI’s current development partners using the tool, nonprofit health system Lifespan’s Norman Prince Neuroscience Institute, is using the technology to help patients regain their voice. For example, the tool was used to restore the voice of a young patient who had lost the ability to speak clearly due to a brain tumor by replicating audio he had previously recorded for a school project, the company’s blog post said. Are listed.

OpenAI’s custom speech models can also translate the speech it generates into different languages. That makes it useful for companies in the audio business like Spotify Technology SA. Spotify is already using the technology in its own pilot program to translate podcasts from popular hosts like Rex Fridman. OpenAI also touted other useful applications of the technology, such as creating a wide range of audio for educational content for children.

In this test program, OpenAI requires partners to agree to its usage policies, obtain consent from the original speaker before using audio, and ensure that the audio you are hearing is generated by AI. is requested to be disclosed to listeners. The company is also installing an inaudible audio watermark to help users distinguish whether audio was created by the company’s tools.

OpenAI said it is seeking feedback from outside experts before deciding whether to release this feature more broadly. “It’s important that people around the world understand where this technology is headed, whether or not we ultimately deploy it widely within our company,” the company said in a blog post.

OpenAI also wrote that it hopes previews of its software will “stimulate the need to strengthen society’s resilience” to the challenges posed by more advanced AI technologies. For example, the company called on banks to phase out voice authentication as a security measure for accessing bank accounts and sensitive information. It also calls for public education about deceptive AI content and further development of technology to detect whether audio content is real or generated by AI.

(Except for the headline, this story has not been edited by NDTV staff and is published from a syndicated feed.)