ChatGPT developer OpenAI unveiled a voice cloning program on Friday. The model called ‘Voice Engine’ can duplicate a human’s voice based on a 15-second audio snippet, according to a blog post.
ChatGPT developer OpenAI unveiled a voice cloning program on Friday. The model called “Voice Engine” can duplicate a human’s voice based on a 15-second audio snippet, according to a blog entry from OpenAI, in which the results of a test with the program are presented.
“We recognize that generating votes that resemble human voices poses serious risks,” which are particularly important to consider in an election year. The San Francisco company works with partners from the areas of politics, media, entertainment, education and civil society, and their feedback is taken into account during development.
Experts fear misuse of applications that use artificial intelligence (AI) this year, which includes European elections and the US presidential election. There are a growing number of voice cloning programs that are cheap, easy to use, and difficult to track.
OpenAI stated that it is aware of these issues. Due to the potential for misuse of synthetic voices, the company has decided to take a “cautious” approach to wider release. Partners who tested “Voice Engine” agreed to rules. These therefore require, among other things, the express consent of every person whose voice is duplicated by the program. In addition, it must be made clear to listeners that the voices they hear were generated by AI.
OpenAI initially caused a sensation with the introduction of ChatGPT in November 2022. The program is able to use artificial intelligence (AI) to generate essays, poems or conversations from very short prompts. ChatGPT suddenly made a large audience aware of the possibilities of AI. At the same time, however, fears about the possible dangers of the technology grew.
The company, which also developed the image generator Dall-e, also introduced an AI application for producing realistic videos called Sora in February.
HOME PAGE