OpenAI is providing restricted access to an AI voice cloning platform named Voice Engine, that is able to produce a synthesized voice based on a 15-second recording of someone’s voice. The AI-generated voice can read out text prompts in various languages, and is currently being tested by companies including Age of Learning, HeyGen, Dimagi, Livox, and Lifespan. OpenAI has been developing Voice Engine since late 2022 and it has already been used to power preset voices for OpenAI’s text-to-speech API and ChatGPT’s Read Aloud feature. The model was trained on a combination of licensed and publicly available data.
AI text-to-audio generation is a growing area of generative AI, and companies such as Podcastle and ElevenLabs are also involved in providing AI voice cloning technology. The U.S. government is working to regulate the use of AI voice technology to prevent unethical uses, such as banning robocalls that use AI voices. Partners who have access to Voice Generation are required to follow usage policies that prohibit the impersonation of people or organizations without consent, and require explicit and informed consent from the original speaker. OpenAI suggests that organizations implement measures to limit the emerging risks caused by AI voice generation technology, such as eliminating voice-based authentication biomarkers for access to bank and other sensitive accounts, and by implementing tracking systems for AI tools and content use.
The whytry.ai article you just read is a brief synopsis; the original article can be found here: Read the Full Article…
