New Voice Deepfake Technology Appears, OpenAI Hasn't Released To Avoid Risks

JAKARTA - OpenAI, a leading technology company, has developed a revolutionary voting cloning software, but chose not to release it to the public due to too high risk potential.

The technology, known as Voice Engine, is capable of replicating anyone's voice in just 15 seconds from the recording of their conversation. Nonetheless, the company considers this technology 'too risky' to be widely used, especially in election years in many countries.

"We recognize that producing a voice that resembles the voice of a person has a serious risk, especially in the election year," said the San Francisco-based OpenAI.

They insist that although they plan to provide an early preview to users, this technology will not be widely released at this point due to the dangers of abuse that may occur.

Investigations are being carried out by authorities in New Hampshire over the thousands of automated calls sent to voters ahead of presidential elections using votes generated by artificial intelligence to emulate President Joe Biden.

Although some startup companies have sold voice cloning technology, OpenAI insists that early users of Voice Engine have agreed not to imitate someone without their consent and to reveal that the resulting sound is artificial intelligence.

OpenAI, known for its chatbots and DALL-E image generators, has similar plans for Voice Engine as they made with the announcement but did not broadly release previous Sora videos.

However, the brand app submitted on March 19 shows that OpenAI is most likely aimed at getting involved in the voice recognition business and digital voice assistants. Thus, such technological upgrades can help OpenAI compete with other voice products like Alexa from Amazon.