OpenAI, the creator of ChatGPT, has unveiled a new artificial intelligence tool called Voice Engine that can create realistic human-like voices. The tool only requires a 15-second audio sample to closely replicate the original voice of the speaker, making it a powerful tool for reading, translating languages, and giving a voice to those who cannot speak.
However, OpenAI is hesitant to fully release Voice Engine due to concerns about deepfake technology. The company acknowledges the serious risks associated with this technology, especially during an election year. They have conducted private testing with trusted partners who have agreed to adhere to strict usage policies, including obtaining explicit consent from the original speaker and disclosing that the voices are AI-generated.
OpenAI is calling for a dialogue on the responsible deployment of synthetic voices and is considering wider release possibilities based on the results of small-scale tests and discussions with stakeholders. They recommend implementing voice authentication to verify the original speaker’s intent and maintaining a “no-go voice list” to prevent creating voices resembling prominent figures.
The caution from OpenAI comes in light of recent incidents involving deepfake technology in politics, such as fake robocalls impersonating President Biden and deepfake videos of politicians. As the technology continues to advance, OpenAI is taking a proactive approach to ensure the responsible use of AI-generated voices in society.