OpenAI, a company known for creating AI that can write texts and create pictures and videos, has now developed a groundbreaking new AI called Voice Engine. This innovative technology has the ability to mimic a person’s voice after listening to just 15 seconds of them talking. Once the voice is copied, Voice Engine can be used to say anything, even in different languages.
Currently, only a select group of companies have been granted access to test Voice Engine. These companies are utilizing the AI to assist individuals who are unable to read, translate content into various languages, and aid those who have lost their ability to speak.
To replicate someone’s voice, Voice Engine only requires a brief recording of their speech and some text to read. The resulting voices are incredibly realistic and can even convey emotions effectively.
OpenAI is proceeding cautiously with the deployment of Voice Engine due to concerns about potential misuse, such as deception or impersonation. With the upcoming election year, the company is particularly wary of fake voices being used to disseminate false information.
The companies testing Voice Engine are required to adhere to strict guidelines. They must obtain permission before replicating someone’s voice and disclose when a voice has been generated by AI. Additionally, OpenAI has implemented special markers on the audio produced by Voice Engine for traceability.
While only a limited number of companies are currently experimenting with Voice Engine, OpenAI is committed to enhancing the technology and ensuring its safe and responsible use. Jeff Harris, a product staff member at OpenAI, expressed the company’s cautious approach, stating, “We’re curious about making it publicly available, but obviously, that comes with added risks in terms of exposure and breaking it.” As OpenAI continues to refine Voice Engine, they are collaborating with government entities and experts to establish best practices for its utilization.