OpenAI aims to start a dialogue on the responsible use of synthetic voices and to make an informed decision about the technology's wider deployment based on these discussions. The company has also called for the phasing out of voice-based authentication for security purposes and the exploration of policies to protect individuals' voices in AI. Despite OpenAI's cautious approach, competitors such as ElevenLabs already offer similar technology to the public, requiring only a few minutes of audio to generate a complete voice clone.
Key takeaways:
- OpenAI has developed a tool, Voice Engine, that can generate a convincing clone of anyone’s voice using just 15 seconds of recorded audio, but has deemed it too risky for general release due to potential misuse in spreading misinformation.
- Voice Engine was first used for the text-to-speech feature in ChatGPT, OpenAI's leading AI tool, and its capabilities have not been publicly revealed due to OpenAI's cautious approach to its release.
- OpenAI hopes to start a dialogue on the responsible deployment of synthetic voices and is conducting small-scale tests to inform decisions about wider deployment of this technology.
- Despite the potential risks, the technology has been used in real-world applications by partners given access to it, including education technology firm Age of Learning and AI visual storytelling app HeyGen, and has even been used to "restore the voice" of a woman who had lost her voice due to a brain tumour.