Copyright Wars Repeat: From Soribada to AI

Around 2000, "Soribada" emerged alongside the spread of high-speed internet, becoming a storm that shook Korea's music industry. Users were thrilled to download free music with a single click without buying tapes or CDs, but the record industry's response was cold. Outrage over "why is our music being used without permission" led to massive lawsuits. Eventually, courts recognized the service operator's liability.

Korea's music market subsequently reorganized around legal streaming services. After legal disputes and various controversies, the common understanding that "fair compensation must be paid to listen to music" became established. Now, with the emergence of artificial intelligence, we are witnessing similar conflicts.

The first battlefront is "AI's use of training data."

AI companies claim they "simply learned from data, like a child reading books." Creators counter that unauthorized learning without permission ultimately threatens their livelihood. Many webtoon and web novel authors I meet in the field are demanding "AI training prohibition" clauses in their contracts, clearly demonstrating this sense of crisis.

Consequently, lawsuits over training data are mounting worldwide. In the United States, authors filed suit against Anthropic, the company operating Claude. The trial court found that learning using illegally downloaded files constituted infringement but acknowledged the possibility of "fair use" for content from directly purchased and scanned physical books. Anthropic made the pragmatic choice to pay substantial settlements to copyright holders, ending disputes early.

In Germany, a music copyright association sued OpenAI, the company operating ChatGPT. The trial court ruled that AI storing and outputting lyrics without rights holders' permission constituted copyright infringement, interpreting "fair use" narrowly. However, OpenAI has appealed, and the dispute continues.

In Korea, the three major terrestrial broadcasters filed suit against Naver in January last year, launching full-scale courtroom battles over training data copyright. The government's recently announced "AI Action Plan" pledged to establish institutional improvements regarding training data by the second quarter of this year, making this year a watershed moment for social consensus on protecting creators' rights.

The second battlefront is whether rights should be recognized for AI-generated outputs.

The current international standard for copyright recognition is "human creative contribution." AI has enabled anyone to easily produce creative works, but paradoxically, the lower human contribution becomes, the more such outputs remain outside legal protection.

This discussion is urgent because it extends beyond technical issues to the future content industry ecosystem. If standards for recognizing copyright in AI-generated works remain ambiguous, companies may hesitate to invest, unable to protect rights to content produced with substantial capital. Conversely, indiscriminately granting copyright to outputs with minimal human involvement risks encroaching on existing creators' domains and disrupting the copyright system itself. Therefore, clear guidelines on "human creative contribution" must precede other efforts to enhance industry predictability.

Major jurisdictions including the United States, EU, and Korea are accelerating the development of AI copyright guidelines. As precedents and cases accumulate, the scope of "human creative contribution" is expected to become increasingly concrete.

The Soribada debate was contentious but ultimately produced the new order of legal streaming. Disputes surrounding AI will follow a similar trajectory as precedents accumulate and legislation and institutional improvements follow. However, given that AI's development pace is incomparably faster than during the Soribada era, regulatory development must also accelerate. Just as Soribada's chaos became the foundation for K-content, I hope current confusion will become the starting point toward a fairer and more sustainable creative ecosystem.