Catching "AI Bamyanggaeng"... Korea Music Performers Association Joins National R&D for Music Copyrights

Blocking Unauthorized Use of Singers' Voices

Image to aid understanding of the article. Provided by Eumsillyeon.

The Korea Music Performers Association announced on January 15 that it will participate as a joint research institution in a national research and development (R&D) project for verifying copyrights of artificial intelligence (AI) and deepfake music.

This research is part of the New Technology Convergence Copyright Technology Development Project, supported by the Ministry of Culture, Sports and Tourism and the Korea Creative Content Agency, and will continue until 2028.

The main objective is to develop core technologies for verifying the copyrights of AI-generated and deepfake music. The Korea Music Performers Association will participate as a joint research institution, collaborating with the lead research institution Electronics and Telecommunications Research Institute (ETRI), as well as the Korea Copyright Commission, MarkAny, OddEye, and Korea University’s Industry-Academic Cooperation Foundation.

Deepfake music refers to audio tracks created by using AI technology to alter a specific singer’s voice or melody so that it sounds as if the original artist performed it. Kim Seungmin, Executive Director of the Korea Music Performers Association, stated, “There is a flood of audio tracks generated without the consent of vocal performers, leading to serious issues of voice rights and publicity rights infringement. We are seeking ways for performers to protect their rights and generate revenue.”

In the first phase of research in 2025, the association established a vocal database (DB) and singing data to identify singers’ audio tracks. During the data collection process, legal and ethical reviews and consent procedures were strengthened to implement safeguards against unauthorized use. During the research period, the association will secure licenses and carry out technology demonstrations necessary for AI technology development.

The research will develop audio watermarking-based rights information embedding technology and an automated deepfake music detection system. Through this, a technological foundation will be established to detect unauthorized use of performers’ voices and track rights information, while providing a transparent settlement and distribution system for performers who have consented to AI use.

Park Keunik, Head of the Business Team at the Korea Music Performers Association, stated, “This technology development, based on our members’ data, will serve as a cornerstone for protecting performers’ rights and building a fair music ecosystem in the AI era.”

Founded in 1988, the Korea Music Performers Association is an organization that collects and distributes neighboring rights royalties for performers in all genres, including popular music, traditional Korean music, and classical music.