Painting Styles and Voices Fall Outside Copyright Protection...
Debate Grows Over AI Imitation

Academia: "Outcome Imitation and Identification Marker Clauses May Apply"

Legal Community: "Risk of Overreach" vs "Need to Expand Portrait Rights"

Artificial intelligence (AI) can easily imitate the brushstrokes of a painter and replicate a singer’s voice. Copyright law cannot prevent this. That is because painting styles and vocal timbres are classified as "ideas" and are therefore not protected. At a recent "Grand Debate on Copyright Issues" hosted by the Ministry of Culture, Sports and Tourism and the Korea Copyright Commission, academia turned its attention to the Unfair Competition Prevention Act as an alternative to fill this gap.


A still cut from the movie <i>How Do You Live?</i> produced by Studio Ghibli.

A still cut from the movie How Do You Live? produced by Studio Ghibli.

View original image

The "Studio Ghibli-style" image craze was what exposed this loophole. Kim Won-O, a professor at Inha University Law School, noted, "Ghibli-like images cannot receive legal protection." He assessed that the current framework—which has regarded styles such as painting or fonts merely as ideas—has reached its limit in an era where generative AI is churning out massive volumes of images. He pointed out that style imitation, AI deepfakes, and AI cover songs are all issues of the same kind.


The issue extends beyond style. Faces and voices are just as vulnerable. Kim Kyung-Sook, a professor of intellectual property at Sangmyung University, pointed out, "The Copyright Act does not adequately protect a particular person's face, voice, name, personal characteristics, or the creator's abstract style itself." She continued, "If a deepfake does not replicate existing photos, videos, or music, copyright infringement may not technically occur, but it can still seriously harm an individual's identity, reputation, privacy, or economic value."


Kim Kwang-Nam, presiding judge at the Seoul High Court, also stated, "The personal markers of certain creators—such as painting style, appearance, voice—are not subject to regulation under copyright law." He also drew a line by stating that other laws currently being discussed, such as the Unfair Competition Prevention Act, are not intended to supplement the limitations of copyright law. "Each law simply protects its own specific legal interest," he said.


Nonetheless, more voices are emphasizing the feasibility of these laws in practice. Lee Cheol-Nam, professor at Chungnam National University Law School, cited Paragraph (pa) of Article 2(1) of the Unfair Competition Prevention Act concerning "imitation of outcomes": "If a particular artist's style is intensively learned so as to disrupt or substitute demand for the original creator in the market, this can serve as a supplementary remedy." Regarding Paragraph (ta) of the same article on "identification markers," he explained, "Unfair competition such as AI cover songs and deepfake advertisements that commercially use, without permission, the name, portrait, or voice of widely known domestic personalities can be regulated under this provision."


Sung Ho Park, Honorary Professor at Hanyang University School of Law / Korea Copyright Commission.

Sung Ho Park, Honorary Professor at Hanyang University School of Law / Korea Copyright Commission.

View original image

Professor Cho Hee-Kyung of Hongik University Law School offered a similar perspective. "Even if an imitation of style is not identical to the original, commercial use that free rides on the creator’s achievement and directly encroaches on that market may fall within the scope of Paragraph (pa)," she explained. She argued that, as with AI cover songs imitating a singer's voice or advertisements synthesizing an actor's face, infringement involving identification markers at the output stage can be directly addressed by the identification marker provision.


Cha Sang-Yuk, professor at Kyungpook National University Law School, discussed cases in which style is imitated or used without authorization, stating, "Even if copyright infringement does not arise due to the idea-expression dichotomy, such acts may still be regulated as unfair competition under Paragraph (pa)." He explained that, when AI recreates a famous singer's voice or produces an advertisement with a celebrity’s face, there can be simultaneous issues with copyright infringement, civil law violations of portrait and publicity rights, and breaches of the Unfair Competition Prevention Act.


Not all cases are covered. Professor Cha stated, "If a new song is created but a singer's voice is learned and commercially imitated by AI, this could be a publicity rights issue under the Unfair Competition Prevention Act." However, he also noted, "A person’s face or voice itself is not necessarily a copyrighted work, so there can be situations where existing copyright law cannot provide a solution."


Legal practitioners expressed concerns on the opposite end of the spectrum about this gap. In particular, lawyers who handle litigation and legal consultation worry about overreach regarding Paragraphs (pa) and (ta). Attorney Kim Hyeong-Ji of Kim & Chang commented about the imitation clause, "This article could leave room for judging even a creator’s unique style, consistent technique, or even unprotected content as illegal." Regarding Paragraph (ta), she stated, "It specifically designates as unfair competition the use, without permission, of identification markers such as name, portrait, or voice that have economic value, thus protecting their economic interests." This means that celebrities’ faces and voices are protected. However, she pointed out, "There are still limitations in protecting the portraits or voices of ordinary people."


Sangyuk Cha, Professor at Kyungpook National University Law School / Korea Copyright Commission.

Sangyuk Cha, Professor at Kyungpook National University Law School / Korea Copyright Commission.

View original image

Attorney Lim Sang-Hyuk of SHIN & KIM instead focused attention on the potential for expansion of portrait rights themselves. He noted that the issue of unauthorized learning and synthesis/distribution of faces and voices included in AI training data is "a question that not only involves copyright law but also collides with the legal reasoning of other statutes." He argued that how far the constitutional and civil approaches to portrait rights, established long before, can be expanded through the advent of AI training is a new and pressing question.


There is also a view assigning considerable weight to the application of the imitation-of-results clause. Choi Seung-Jae, professor of law at Sejong University, stated, "Even if a particular creator’s unique painting style is intentionally and intensively learned to launch a similar commercial service, this can be construed as a violation of the Unfair Competition Prevention and Trade Secret Protection Act, thereby protecting the creator’s economic interests."


Some voices raised even more fundamental issues. Sungho Park, Emeritus Professor at Hanyang University School of Law, stated, "If human creative contribution is not recognized in the result produced by AI, it belongs to the public domain and can be freely used by anyone." He emphasized, "The Paragraph (pa) provision of the Unfair Competition Prevention Act should not be applied lightly to such AI outputs."


Sukkwon Kang, Chairman of the Korea Copyright Commission / Korea Copyright Commission.

Sukkwon Kang, Chairman of the Korea Copyright Commission / Korea Copyright Commission.

View original image

Professor Cho Hee-Kyung also cautioned, "If Paragraphs (pa) and (ta) are overused, they may end up indirectly protecting ideas or expired works that copyright law intentionally chose not to cover, so prudent judgment is required."



Ultimately, academia agrees that copyright law alone cannot address the problems of style or voice imitation. What’s needed is a multilayered framework where copyright law protects forms of expression, and the Unfair Competition Prevention Act protects data, achievements, and personal identification markers. However, precisely where the boundary should be drawn remains a matter of ongoing debate.


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing