From Ziff Davis: 2024-03-30 12:10:25

OpenAI introduced voice capabilities in ChatGPT last September. The company is now showcasing Voice Engine, which can generate longer audio from a 15-second audio clip and text prompt, producing lifelike voices with inflection. Voice Engine has been tested for various scenarios, emphasizing safety and building partnerships in specific industries.

Despite concerns about the misuse of AI-generated voices, OpenAI reassures the safety of Voice Engine. The company is engaging with governmental, media, entertainment, and educational partners to gather feedback in the development process. While Voice Engine has potential risks, OpenAI is committed to safety, security, and trust, limiting its wide release for now.

OpenAI aims to address challenges posed by convincing generative models like Voice Engine by advocating for a phase-out of voice authentication, policies to protect voices in AI, and technology to detect inauthentic voices. The company stresses societal resilience and the need for safeguards against potential misuse of voice technology in various sectors.



Read more at Ziff Davis: OpenAI previews Voice Engine, synthetic voice creator