OpenAI claims to be developing AI that can mimic human speech

On Friday, OpenAI unveiled samples from initial trials of its latest tool, named Voice Engine, designed to replicate someone's voice convincingly using a 15-second audio clip. With this technology, users can input a paragraph of text, and the tool will articulate it in the AI-generated voice.

While there are already various AI-powered voice services accessible to the public, OpenAI has demonstrated exceptional proficiency, akin to its success with the breakthrough chatbot ChatGPT, in gaining widespread adoption of AI applications.

According to the company, an AI-driven text-to-voice tool could serve multiple purposes such as aiding in translation, assisting children with reading, or facilitating communication for individuals who have lost their ability to speak. However, there are concerns among skeptics that it might also contribute to the spread of disinformation or make perpetrating scams easier.

OpenAI has restricted the use of Voice Engine to a select "small group of trusted partners," encompassing educational and health technology firms. It intends to evaluate feedback from these partners to determine the feasibility and potential risks associated with broader utilization.

In a blog post, OpenAI acknowledged the significant risks involved in generating speech that closely resembles human voices, particularly in light of an election year. Although the company recognizes the necessity for substantial adjustments as AI-generated audio becomes more prevalent, it has no immediate plans to release Voice Engine to the general public. As a precautionary measure, it suggested phasing out voice-based authentication for bank accounts, among other potential actions.