"Use of AI Training Works Is Not 'Fair Use'... Royalties Must Be Paid"

Yoon Doo-hyun, People Power Party Lawmaker - Onshinhyeop Joint Forum
"Fair Compensation Calculation Method Should Be Reflected in Terms and Conditions"

Participants, including Representative Yoon Doo-hyun (fourth from the left in the photo), are taking a commemorative photo on the 5th at the National Assembly Members' Office Building in Yeouido before the discussion forum titled "Generative AI News Content Learning: Issues and Challenges."
[Photo by Representative Yoon Doo-hyun's Office]

Concerns have been raised that the use of copyrighted works in the data training process of generative artificial intelligence (AI) cannot be regarded as 'fair use' under copyright law. Fair use means that reproduction or use without the copyright holder's permission does not constitute copyright infringement if it serves the public interest, but current copyright law provisions cannot be applied to the data training process of generative AI.

Attorney Shim Min-seon (Barun Law LLC) attended the discussion forum titled 'Generative AI News Content Training: Issues and Tasks' held on the 5th at the National Assembly Members' Office Building in Yeouido, Seoul, and argued this point, emphasizing, “Just as copyright property rights are limited on the condition of compensation payment for the use of works in educational processes, it is necessary to introduce related regulations for the use of works in generative AI data training.”

The forum was co-hosted by Representative Yoon Doo-hyun, chairman of the People Power Party’s Media Policy Coordination Special Committee, and the Korea Online Newspaper Association (President Park Hak-yong).

Panelists Director Noh Chang-hee of the Digital Industry Policy Institute and Attorney Kim Jin-wook (Juwon Law LLC) expressed the same view. Director Noh said, “It is difficult to consider the use of copyrighted works in the data training process of generative AI as fair use under copyright law, so legislative alternatives are needed.” Attorney Kim presented overseas cases from Europe, the United States, and Australia related to the valuation of news content and advised, “A detailed approach, including reflecting terms and conditions, is necessary to establish a fair compensation method for news content.”

Shin Han-soo, deputy editor of Seoul Economic Daily, who attended the discussion representing the media industry, emphasized, “A more proactive attitude from AI technology companies is needed to establish standards for the use of news as training data and an appropriate compensation system, thereby creating an environment for producing high-quality news content and fostering a virtuous cycle of generative AI technology development.”

Reporter Kim Tae-gyun of Yonhap News quoted a statement by a German media CEO who described a situation where AI machine learning data can no longer be scraped from the internet as ‘Artificial Stupidity,’ saying, “Platforms should show a proactive attitude in establishing a system that guarantees fair compensation and a ‘win-win’ outcome.”

In response, the AI technology industry took a cautious stance. Ha Jung-woo, head of the AI Innovation Center at Naver Cloud, argued, “If the generative AI ecosystem becomes dependent on global big tech companies, copyright compensation could become even more difficult,” and stressed the need to take time to reach a consensus.

Representative Yoon Doo-hyun, who hosted the forum, stated in his opening remarks, “There are doubts about AI learning news articles written through the efforts and hard work of news organizations and journalists without any compensation,” while Park Hak-yong, president of the Korea Online Newspaper Association, said, “We need to reasonably establish AI news content usage standards and seek a path of coexistence between news copyright holders and AI technology companies.”

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