Korea’s Deepbrain AI files patent to detect fake voices

2024.02.06 13:24:01 | 2024.02.06 18:04:08

[Courtesy of Deepbrain AI]이미지 확대

[Courtesy of Deepbrain AI]

Deepbrain AI, a South Korean generative artificial intelligence (AI) company, has filed a patent for deep voice detection technology based on deep learning, the company said on Monday.

Deep voice refers to fake voices manipulated by AI using deep learning-based voice extraction technology. It is often used for creating fake videos and images.

The technology for which the patent has been filed differs from existing deep voice detection technology in that it considers voice frequency and time to determine whether the voice is synthesized and generated by AI. The plan is to further enhance the solution to enable real-time detection of voice authenticity even during conversations.

Deepbrain AI has recently announced a deepfake detection solution utilizing this patented technology, with an aim to lead in the prevention of AI abuse crimes, including voice phishing.

By Kim Dong-eun and Chang Iou-chung

[ⓒ Pulse by Maeil Business Newspaper & mk.co.kr, All rights reserved]