0
/ Joseon Hip-hop "Forest of the Mind" YouTube

The era of the "solo producer" has arrived, where typing a few sentences generates a complete song you want to hear. As generative artificial intelligence (AI) rapidly integrates into music production, making music has become far easier, yet questions surrounding copyright are also mounting. If a human selects and refines melodies created by AI, to what extent can this be regarded as human creation?

The current copyright system fundamentally protects human creation. While elements creatively modified by a human in works utilizing generative AI can be protected, clear criteria on how much human intervention is required to qualify as creation have yet to be established.

The music industry is also searching for this boundary. The Korea Music Copyright Association (KOMCA) last month permitted trust registration for AI-assisted music in which humans substantially participated in songwriting, composing, or arranging, only to withdraw the decision 22 days later. Since then, the Ministry of Culture, Sports and Tourism has stepped in to discuss standards for copyright trust and royalty distribution for AI-assisted music, moving the discourse to the governmental level.

AI-assisted music has already entered domestic music charts. "Forest of the Mind," released in May by Joseon Hip-hop, peaked at 89th on the Melon Top 100 and reached 99th on the daily chart as of the 29th. For this track, Joseon Hip-hop wrote the lyrics directly and inputted prompts into the generative music AI Suno to handle composition and vocals.

The use of AI is also growing across production settings. Singers and producers who directly produce or take part in writing and composing utilize AI to flesh out ideas or compare multiple melodic and arranging directions. An artists and repertoire (A&R) official stated, "Rather than having AI generate entire songs from start to finish, it is frequently used to verify ideas or weigh different directions," adding that "it offers high efficiency in reducing working hours and enabling diverse attempts in a short span of time."

Another industry insider remarked, "AI has become a tool assisting composers' creative work, and the role of creators is shifting toward planning and setting the direction for what sound to realize," noting that "standards are needed to define how far human planning and directing can be recognized as creation."

The methods of utilizing AI vary widely as well. A person might write the lyrics and melody before adding accompaniment via AI, or select and modify melodies generated by AI. Some use AI strictly for mixing or mastering. Even within the category of AI-assisted music, the role played by the human differs. Consequently, opinions suggest that attention should focus on the creative process rather than just the final output—verifying what the human produced directly, at which stage AI was employed, and how much the final product was modified or restructured.

A music industry insider said, "Rather than the mere fact that AI was used, it should be possible to verify at which stage and in what manner it was utilized," adding, "Only when creators' self-reporting is accompanied by documentation of the working process will practical copyright determinations become feasible."

The music industry is also preparing technology to detect whether AI was utilized. Music-related organizations, including KOMCA, decided to jointly introduce detection software to identify AI-generated and AI-assisted music. The goal is to verify the use of AI from actual audio files rather than relying solely on creators' declarations.

Royalty distribution presents another challenge. If rights are recognized for portions created by humans, assessing the degree of contribution inevitably follows. The impact on existing distribution structures must also be examined if AI-assisted music grows rapidly. How to protect the rights of original copyright holders during the process where AI trains on existing music constitutes a separate issue. While the issue of AI training on existing music and the copyright issue of outputs created using AI are linked, they are not the same. Rather than drawing lines solely based on whether AI was used, criteria are required to verify what processes the creator executed directly and to what extent they intervened in the output.

Pop culture critic Park Song-ah said, "The key lies not in whether AI was used, but in examining how substantially a human intervened in the track's direction, melody, lyrics, and structure," adding, "How much the final product was modified or restructured, and whether that process can be verified through working project files or revision histories, will serve as crucial judgment criteria."

                                                                                                              Lee Da-hye