OpenAI has developed an AI model dubbed ‘Voice Engine’ that exceptionally synthesizes human voice using a brief, 15-second audio sample. This technology contributes to a myriad of sectors, including education, healthcare, and AI communication, and continues to transform how humans and machines interact.
Takeaways:
- Impressive Voice Cloning: This powerful technology designed by OpenAI can clone any voice using a 15-second clip. Upon receiving textual inputs, the AI-generated voice impeccably pronounces them in the same language as the sampled speaker or in other languages.
- Potent Applications: Since its inception in late 2022, the Voice Engine has been instrumental in powering features like Read Aloud in ChatGPT. It has also been adopted by firms such as Age of Learning, HeyGen, Dimagi, Livox, and Lifespan, demonstrating its wide-ranging utility and adaptability.
- Responsibly Managed: OpenAI has set clear partner guidelines that prohibit using Voice Generation to impersonate individuals or entities without consent, and require disclosure that the voices are AI-generated. Additionally, OpenAI uses watermarking to trace the origin of its audio clips and monitor their usage. It further suggests steps to mitigate associated risks, like phasing out voice-based authentication for bank accounts and developing AI content tracking systems.