Nana Wu
Seoul, South Korea – The South Korean music industry has mobilized in an unprecedented display of unity against the escalating challenges posed by artificial intelligence. On February 26, six leading music rights organizations formally launched the K-Music Rights Organization Mutual Growth Committee in Seoul, issuing a stark warning that the next two years are critical to the survival of Korean creators amidst the rapid advancement of AI. This collective action underscores a growing global concern within creative industries about the unchecked proliferation and commercialization of AI-generated content.
The AI Tsunami: A Global Challenge, Korean Urgency
Committee chair Lee Si-ha articulated the gravity of the situation at the launch event, stating, “The next two years are the golden time that will decide the life or death of Korea’s music industry. Individual responses from separate organizations can’t stop this massive wave of change. The entire industry must stand together.” This declaration reflects a shared sentiment among artists, producers, and rights holders worldwide: the pace of AI innovation has outstripped existing legal and ethical frameworks, creating a precarious landscape for human creativity. The music industry, with its intricate web of copyrights, performance rights, and intellectual property, finds itself on the front lines of this technological revolution. Generative AI, capable of composing melodies, writing lyrics, and cloning voices with alarming realism, threatens to fundamentally alter how music is created, distributed, and monetized.
Foundational Unity: The Coalition’s Composition and Demands
The newly formed coalition represents a formidable alliance, bringing together virtually every significant stakeholder in Korea’s vibrant domestic music ecosystem. Its members include the Korea Music Copyright Association (KOMCA), the Korea Music Content Association, the Korea Music Performers Federation, the Korea Recording Industry Association, the Korea Entertainment Producers Association, and the Together Music Copyright Association. This diverse representation — encompassing composers, lyricists, performers, record labels, and producers — ensures a holistic approach to addressing AI’s multifaceted impact.
At its inception, the committee adopted an "AI-Era Music Rights Declaration," outlining three core demands designed to safeguard human creativity and ensure fair compensation:
- A ban on AI training without creator consent: This demand targets the fundamental issue of data sourcing, asserting that existing copyrighted works should not be used to train AI models without explicit permission from rights holders.
- Mandatory transparency in AI generation processes: This calls for clear disclosure mechanisms, allowing consumers and industry professionals to identify whether a piece of music has been entirely or partially generated by AI.
- Clear legal distinctions between human-created and AI-generated works: This aims to establish a legal framework that differentiates between human artistic expression and algorithmic output, potentially influencing copyright eligibility and royalty distribution.
These demands echo similar calls from artists and industry bodies in other regions, highlighting a global consensus on the need for ethical AI development and robust regulatory oversight.
The Precedent: Korea’s First AI Copyright Battle (EvoM Case)
Korea’s proactive stance is rooted in firsthand experience with the disruptive potential of AI in music. The industry was rocked in July 2022 by the "EvoM case," involving trot singer Hong Jin-young’s hit song "Love Is 24 Hours." KOMCA discovered that the song was composed by EvoM, an AI program developed by GIST professor Ahn Chang-wook. EvoM had, over six years, generated an astonishing 300,000 compositions, selling 30,000 tracks and accumulating 600 million won (approximately $450,000 USD) in revenue.
This discovery prompted KOMCA to freeze royalty payments for the AI-created songs, a pivotal moment that ignited Korea’s AI music rights debate. The legal reasoning was stark: Korea’s Copyright Act explicitly defines creative works as “creations expressing human thoughts or emotions.” If an AI is the sole creator, without demonstrable human input reflecting human thought or emotion, there is no legal basis for copyright protection or subsequent royalty payments under existing law. The EvoM case starkly exposed a fundamental legal gap: AI was already producing and commercializing music on a significant scale, yet the legal framework designed for human creators had not evolved to address this new reality. This incident served as a powerful catalyst, propelling the Korean music industry to confront the issue head-on and seek collective solutions.
K-pop’s Unique Vulnerability: Voice Cloning and Deepfakes
Beyond composition, the rapid advancement of generative AI poses an even more insidious threat to K-pop artists: voice cloning and deepfake content. A significant legal gap compounds this problem: under current Korean law, a singer’s unique voice itself is not explicitly defined as a copyrightable work. This omission leaves a critical vulnerability, as AI cover creators can illegally clone the voices of famous K-pop idols, producing unauthorized content that floods online platforms. Existing performer rights protections, designed for traditional forms of infringement, struggle to effectively combat this new wave of synthetic content.
The global exposure achieved by K-pop powerhouses like BTS, NewJeans, and BLACKPINK, while propelling them to international stardom, has paradoxically made them prime targets for AI-generated fake content. A disturbing 2023 Security Hero report revealed that Korean singers and actresses constitute a staggering 53% of individuals featured in deepfake pornographic content worldwide, with eight of the top ten individual targets being Korean female singers. While this specific issue extends beyond music, it highlights the severe ethical and personal violations enabled by advanced AI technology when applied to public figures, especially within the K-pop sphere.
The artistic integrity of K-pop is also under assault. Voice synthesis technology has advanced to a point where fans themselves express confusion, stating they “can’t tell who’s real anymore” when AI cover songs featuring cloned idol voices saturate platforms like YouTube. This erosion of authenticity not only devalues the original artistry but also creates a breeding ground for misinformation and psychological distress for the artists themselves. The emotional and financial toll of such unauthorized voice cloning and deepfake generation is immense, urging a swift and comprehensive legal response.
Industry Response: Internalization and Proactive Policies
In response to these threats, major entertainment companies in Korea are taking strategic measures. HYBE, the powerhouse behind BTS, acquired AI voice startup Supertone for 45 billion won (approximately $33 million USD), securing a 56.1% controlling stake. This move signals a clear strategy: rather than merely waiting for regulations to catch up, Korea’s biggest entertainment players are internalizing AI voice technology, aiming to control its development and application within their own ecosystems. This could lead to authorized and ethical uses of AI voices for various projects, while simultaneously protecting their artists’ vocal identities from unauthorized exploitation.
KOMCA, in particular, has moved faster than many of its global counterparts in establishing a policy framework. As of March 24, 2025, all new music registrations require a signed statement certifying that “AI was not used and the work consists solely of human creative contributions.” False statements can trigger severe consequences, including legal liability, royalty freezes, and the removal of works from the database. This pioneering policy, while strict, does not outright ban all AI use. KOMCA’s official position clarifies that works created “100 percent by AI” cannot be registered. However, if AI serves as an assistive tool – for example, in generating initial ideas or refining arrangements – while the human creator’s core contribution to the “thoughts or emotions” remains clear, copyright protection may still apply. This nuanced approach aligns with guidance from the World Intellectual Property Organization (WIPO), which noted in a 2024 report that “AI-centric creations are difficult to protect under current copyright frameworks” but acknowledges the potential for AI to assist human creativity.
Global Landscape: Litigation, Negotiation, and Licensing
The Korean industry’s battle mirrors a broader global struggle within the music sector. Initially, the global music industry’s approach to AI was characterized by aggressive litigation. In June 2024, the three major labels — Universal Music Group (UMG), Warner Music Group (WMG), and Sony Music Entertainment (SME) — jointly sued AI music startups Udio and Suno. The lawsuits alleged that both platforms had trained their generative AI models on vast quantities of copyrighted recordings without obtaining permission or licensing agreements from the original rights holders. This legal action underscored the industry’s determination to protect its intellectual property and establish the principle that AI developers must compensate creators for the use of their works as training data.
However, through late 2025, the landscape began to shift from purely adversarial courtroom battles towards negotiation and licensing. The major labels reached varying settlements and licensing arrangements with these and other AI music platforms. This evolution signals a pragmatic recognition within the industry that coexistence with AI is inevitable. Instead of attempting to halt AI development entirely, major labels are now seeking to establish structured commercial relationships with AI companies, aiming to monetize the use of their extensive catalogs for AI training and ensure a share in the economic value generated by AI-created music. While the legal conclusions regarding AI’s copyright status remain largely unwritten, this shift marks a significant step towards integrating AI into the existing music economy, albeit under terms that protect creator rights.
The Path Forward: Blockchain for Transparency and Tracking
Looking ahead, the newly formed K-Music Rights Organization Mutual Growth Committee has ambitious plans to build a blockchain-based unified infrastructure. This innovative system aims to track AI-generated and distributed music data with unprecedented transparency and accountability. The envisioned infrastructure would connect international standard identification codes (such as ISWC for musical works and ISRC for sound recordings) with content identification systems used by major streaming platforms and digital service providers.
The primary goal is to create auditable records of AI training pathways. This means the system would ideally be able to trace which existing copyrighted works were used to train a particular AI model, how an AI-generated track was created, and its subsequent distribution. Such a robust data structure would significantly enhance copyright tracking, making it easier to identify infringement, ensure proper attribution, and facilitate royalty collection for creators whose works contribute to AI-generated content. By leveraging blockchain technology, the committee hopes to establish an immutable and transparent ledger of AI music creation and usage, providing a crucial tool for governance in an increasingly complex digital landscape.
Challenges and the Race Against Time
The question now is whether these proactive measures and ambitious technological solutions can match the relentless pace of AI development and the complexities of regulatory reform. Korea has already experienced the direct cost of legal gaps through the EvoM case, demonstrating the urgent need for updated legislation. KOMCA’s pioneering policy and the solidarity committee’s launch position Korea among the fastest-moving nations in establishing governance frameworks for AI in music.
However, significant structural gaps persist. These include:
- Inadequate legal protections for vocal identity: As highlighted by the K-pop deepfake crisis, current laws do not sufficiently protect an individual’s unique vocal signature from AI cloning.
- Unclear standards for determining copyrightability of AI-created works: The distinction between AI as an assistive tool and AI as the sole creator remains legally ambiguous, particularly as AI models become more sophisticated.
- Limited enforcement mechanisms: Against the backdrop of a vast internet, effectively policing platforms that host unauthorized AI covers and deepfakes presents a monumental challenge.
Whether this industry-wide coalition can move beyond declarations to meaningful institutional reform and implement effective technical defenses will become clear within the next two years. The stakes are incredibly high, not just for the Korean music industry, but as a potential model for other creative sectors globally grappling with the same existential questions. The countdown has begun to define the future of music in an AI-driven world, balancing innovation with the protection of human artistry.
