go to news

detail

* It has been translated through NAVER
papago logo

“AI Learning News for Free? Standards for Compensation Must Be Established”

PARK Yoonyea
Input : 
2024-11-26 15:32:04
Updated : 
2024-11-28 15:07:02
Comprehensive Discussion at the ‘AI Era News Copyright Forum’ The article is quality learning material Fair use is a contentious issue Differential application of compensation standards Promoting coexistence between AI companies and media
Participants are presenting their opinions at the comprehensive discussion held by the Korea Press Foundation on the 25th. <Korea Press Foundation>
Participants are presenting their opinions at the comprehensive discussion held by the Korea Press Foundation on the 25th. <Korea Press Foundation>

“It is difficult to consider it fair use when AI developers replicate articles and provide them as AI responses.” (Lawyer Yang Jin-young)

“The compensation for using articles can vary depending on whether the use of AI's learning data is for commercial or non-profit purposes.” (Lawyer Shin Yong-woo)

In the age of artificial intelligence (AI), the media industry is seeking solutions to protect news copyright. On the 25th, the Korea Press Foundation and six media organizations held a comprehensive discussion at the Press Center in Jung-gu, Seoul, titled ‘AI Era News Copyright Forum.’ Since the forum was launched in March, media organizations have been discussing responses to changes in the digital environment, including news copyright protection and generative AI technology, for nine months.

The comprehensive discussion was held to share the results of this forum and engage in related discussions. Lawyers Yang Jin-young from the law firm Minhoo and Shin Yong-woo from the law firm Jipyeong, who served as committee members of the forum, presented the results.

In the AI era, newspaper articles are quality materials for AI developers when training AI. As the AI search market expands, the commercial value of the latest information will increase.

When AI learns from news content, the issue of copyright infringement is unclear, leading to numerous lawsuits between AI developers and media companies in the United States. The lawsuit between The New York Times and OpenAI is a representative case. In South Korea, no lawsuits have been filed yet.

Whether the use of news content for AI development can be considered fair use is crucial. Lawyer Yang Jin-young stated, “It is difficult to qualify as fair use,” adding that “the transformative use of learning data and its commercial nature are important factors in determining fair use, and there is a risk of copyright infringement.” He explained that the outputs of generative AI using news articles have the same or very similar purpose of providing current information as the original news articles, making it difficult to meet the requirements for transformative use.

The commercial or profit-oriented use by AI developers is also difficult to recognize as fair use. However, there are claims that it could be exceptionally recognized as fair use for public reasons such as research or education. Nevertheless, Lawyer Yang countered, “In the development of generative AI, which involves substantial technology and budget, it is challenging to clearly distinguish between profit-oriented and non-profit characteristics.”

The text and data mining (TDM) exception clause is necessary for fostering the AI industry, but the rights of copyright holders (media companies) to refuse data access must also be guaranteed. The TDM exception clause allows the use of copyrighted content for analysis without the copyright holder's permission for research or scientific purposes. The European Union has legislated this clause, while South Korea has not yet done so.

Lawyer Yang stated, “If the TDM exception clause is legislated, an essential condition is to recognize the right of copyright holders to reserve data access and to ensure lawful access.”

Discussions were also held on the standards for determining compensation for the use of news content. Lawyer Shin Yong-woo stated, “The standards for determining compensation can vary based on the purpose of use, the size of the company, the recency of the content, and the contract duration,” adding, “The price can vary depending on whether the use of AI learning data is for commercial purposes, the size of the AI company, and whether the news is recent. In a situation where it is still difficult to gauge the AI industry, it is easier to set the contract duration short, on a yearly basis.”

Coexistence and cooperation between AI developers and media companies were also emphasized. Professor Lee Dae-hee from Korea University Law School stated, “Recently, AI developers are using RAG (Retrieval-Augmented Generation) technology to clarify sources such as articles to prevent AI from spreading misinformation,” adding, “This aspect could be a point of coexistence between AI developers and media companies.”