New OpenAI tool is capable of “cloning” human voices

OpenAI, creator of ChatGPT, shared this Friday (29 March 2024) the first results of a test for a generative AI (artificial intelligence) resource that can read words in a voice very similar to that of a human. The Voice Engine allows the user to enter text and send a 15-second audio sample for the information to be reproduced in the suggested voice.

The company has been developing the tool since 2022 and is in the testing phase. However, OpenAI says it adopts a “cautious approach” regarding the resource and has not yet defined whether the tool will be released to the general public.

“We are taking a cautious and informed approach to wider dissemination due to the potential for misuse of synthetic voice. […] Based on these conversations and the results of these small-scale tests, we will make a more informed decision about whether and how to implement this technology in [larga] scale”the company said in a publication.

“We believe that any broad deployment of synthetic voice technology must be accompanied by voice authentication experiments that verify that the original speaker is knowingly adding their voice to the service, and a banned voice list that detects and prevents the creation of voices that are too similar to prominent figures”, declared.

The North American company also highlighted the care taken in launching the Voice Engine during the election year in the United States. “We recognize that generating speech that resembles people’s voices carries serious risks, which are especially important in an election year.”