OpenAI has announced a limited-scale preview for its latest tool, Voice Engine, which utilizes voice cloning technology to replicate any speaker’s voice with remarkable accuracy. By analyzing just a 15-second audio snippet, Voice Engine can generate speech that sounds natural, complete with emotive and realistic tones.
The development of Voice Engine has been ongoing since 2022 and is built upon OpenAI’s existing text-to-speech API. Samples showcased on the company’s official blog demonstrate the tool’s ability to produce authentic renditions of human voices, showcasing its potential applications in reading assistance, language translation, and aiding individuals with disabilities.
Also Read: OpenAI’s “Accidental” Leak Reveals ChatGPT’s Next Major Upgrade
One notable example highlighted by OpenAI involves a patient with speech impairment who benefited from Voice Engine’s capabilities. By using a clone generated from audio recorded during a school project, the individual was able to utilize Voice Engine through a pilot program at Brown University.
However, alongside its potential benefits, Voice Engine also raises concerns regarding misuse by bad actors. The technology could be exploited to create highly realistic deepfakes, posing significant privacy and security risks. OpenAI acknowledges these challenges, particularly in the context of an election year, and is actively soliciting feedback from various stakeholders to address these concerns.
As of now, OpenAI has not provided a release date for Voice Engine, indicating that further refinement and consideration of privacy issues are necessary before a public release.
