OpenAI has unveiled a groundbreaking new tool called Voice Engine, capable of transforming any 15-second voice sample into an exact synthetic replica.
Voice Engine's potential extends beyond mere convenience – it offers significant aids to those unable to read by vocalizing text in a familiar voice. Additionally, it promises to break down language barriers, allowing content creators to reach a global audience in their authentic vocal style. Moreover, it provides a voice to those who have lost theirs due to medical conditions or other reasons.
Voice Engine is part of OpenAI’s broader suite of AI-driven innovations, integrated within platforms like ChatGPT. The company first developed Voice Engine in 2022 and has used it to power the preset voices available in the text-to-speech API as well as ChatGPT Voice and Read Aloud.
According to Jeff Harris of OpenAI in his interview to TechCrunch, while the technology holds immense promise, its deployment is approached with caution to mitigate potential misuse.
This responsible development approach extends to data privacy and ethical concerns. OpenAI ensures that Voice Engine is trained on a combination of licensed and publicly available data, with strict policies against using personal data without consent.
