Korean Newspaper Association to Sue AI Companies Over Unauthorized Use of News Content

On February 17, the Korean Newspaper Association announced plans to report companies that unlawfully utilize news content from media outlets for training generative artificial intelligence (AI) to the Fair Trade Commission. The association stated that it will first file a lawsuit against Naver, which has acknowledged the unauthorized use of its generative AI services, HyperCLOVA and HyperCLOVA X. Additionally, it plans to gradually pursue legal action against international AI companies such as OpenAI and Google, which are also suspected of misusing news articles. The association aims to establish a fair trading relationship between newspapers and generative AI companies through this legal action, while also seeking to ensure that the value of news content is recognized and that the algorithms of IT platforms are transparent.

The association has formed a task force consisting of ten digital strategy leaders from member companies to address the issue of unauthorized use of news content by generative AI firms. They have been examining the legality and unfairness of these practices, concluding that the unauthorized use of news content for AI training violates copyright laws and constitutes abuse of market dominance and superior bargaining position as prohibited by fair trade laws. The association criticized these companies for failing to compensate media outlets for the use of news in training their large language models (LLMs) and AI search services, as well as for replicating article content without proper attribution. They also highlighted the lack of transparency in AI algorithms related to news article placement and the unilateral changes to contracts regarding the use of news content as unfair practices.

Furthermore, the association is advocating for legal and institutional reforms to address copyright infringements by generative AI. They plan to request amendments to the 'Basic Law on the Advancement of Artificial Intelligence and the Establishment of Trust,' which would mandate the disclosure of sources for AI training data. They also aim to remove Article 7, Clause 5 of the copyright law, which currently states that 'news reports that merely convey facts' are not protected by copyright, and to add news as a separate category for copyright protection.