Korea's AI Pop Hits and the Copyright Law That Can't Keep Up

The No. 3 Track Nobody Sang

The song is called “꿈꾸던 어른이 되었나요” — “Did You Become the Adult You Dreamed Of?” — and its opening lyrics about the quiet weight of grown-up worries seem tailor-made for a post-work commute. The voice is warm, human, convincing. It is also entirely synthetic.

The artist is listed as 머무르 (Meomuru). According to the producer, the human contribution was limited to planning the project and writing the lyrics. The composition and vocals were generated by artificial intelligence. As of Oct. 13, the track sat at No. 3 on Melon, South Korea’s largest music streaming platform, marking the first time an AI-composed and AI-vocalized song has reached that height on a Korean chart. Its music video has crossed 10 million YouTube views.

This is not a novelty act. It is a canary.

Across Korea’s hyper-digitized music market — the world’s most idol-driven and IP-commercialized ecosystem — AI-generated tracks are multiplying at a pace that outstrips every existing legal framework. The result is a regulatory vacuum that threatens not only songwriters and composers, but the entire architecture of K-pop’s revenue model. And the consequences will not stop at Korea’s border.

The Speed Problem

The technology is moving faster than the law in ways that make enforcement nearly impossible. Suno, one of the leading generative AI music platforms, has iterated repeatedly and now allows any user to produce a professional-quality track in one or two minutes. The output is indistinguishable from work created by human composers — to trained ears included.

A second AI-generated track, “마음의 숲” by the group 조선힙합, reached No. 17 on YouTube Music’s chart on the same day. On Deser, a French streaming platform, an average of 90,000 AI-generated tracks were produced per day in June alone. The flood is global; the legal response is nil.

The practical reality on the ground is already stark. An anonymous Korean composer told the press that more than 80 percent of working songwriters use AI in some form — large or small. When usage is this widespread and undetectable without self-reporting, any copyright system that relies on human creators declaring their tools is fundamentally broken.

The Regulatory Vacuum

In August, the Korea Music Copyright Association (음저협) proposed a narrow criterion: it would allow collective-management registration for parts of a song where human contribution could be clearly identified — lyric writing, composition, arrangement — even if AI tools were involved in the process. The plan was immediately pulled after criticism from the National Assembly and the government that broader social consensus was needed before any formal standard could take shape.

Until a standard is established, the association has opted for a blanket pause: no AI-generated recordings will be registered for copyright protection. This leaves a legal gray zone that is actively being exploited.

The difficulty is not merely ideological — it is technical. There is currently no reliable method to detect whether a track was AI-generated unless the creator self-reports. Even dedicated AI-detection software would fail to catch cases where a composer borrowed only a melody or a creative idea from an AI output, rather than copying the full generated file. A system built on detection technology is therefore unreliable by design.

The K-Pop IP Question

Why does this matter beyond a few viral AI pop songs? Because K-pop is one of the most tightly controlled intellectual-property systems in global entertainment. A single hit track is a revenue node that feeds management companies, publishing arms, performance rights organizations, and international licensing deals. The model depends on clear attribution, enforceable rights, and scarce creative output that commands premium pricing.

AI-generated music flips that scarcity model on its head. When anyone can produce a radio-ready track in two minutes, the economic value of a composed melody collapses. If the current trajectory holds, the professional songwriting workforce — already under pressure from streaming economics — faces a deeper structural threat.

Industry veterans warn that if Korea’s creative environment continues to contract while AI music proliferates unchecked, the country’s competitive edge in K-pop will erode. The logic is straightforward: an industry that cannot monetize its core creative assets loses its ability to fund the next generation of artists.

A Path Forward

Some experts argue that the analogy to sampling is instructive. Decades ago, borrowing segments of existing recordings was treated as plagiarism and legal risk. Today, sampling is an accepted creative technique with established licensing frameworks. The question is whether generative AI use in composition will follow a similar arc — from taboo to regulated norm.

One proposal gaining traction is a negative-regulation approach: rather than banning AI-generated music outright, the system would permit it with a mandatory disclosure requirement. Creators would be required to declare whether AI was used in the composition or vocal production. This shifts the burden from impossible detection to verifiable self-reporting, while preserving the creative output that the market clearly demands.

The Korean case matters globally because it sits at the intersection of three forces: a market that is more digitized than almost any other, an idol industry whose IP model is the most aggressively monetized in the world, and a regulatory environment that is still searching for a consensus. Japan and China, both major music markets with their own AI adoption curves and regulatory debates, will be watching closely. So will the major labels and streaming platforms that are already adjusting their content policies in real time.

The No. 3 track on Melon is not the future of music. It is the present. And the law has not yet caught up.