[Veat's Startup Legal Walk] Is AI Learning Data Collection Legitimate?

Would it be okay to use data found on the internet?

Originally, a person’s creative works (such as writing, photos, paintings, songs, movies, plays, books, and computer software) created with their own ideas or feelings are considered “works” and are protected by copyright. The person who creates a work has the exclusive right to use it and the right to allow others to use their work. Using a work without the permission of the copyright holder is a violation of copyright law. 

This is why the question of whether obtaining high-quality data from the internet for creating high-level AI models is a recent subject of debate. Weather statistics from Seoul and legal documents are not considered works, so they can be easily used for AI model learning, but newspaper articles, discussions posted on forums, comments, and photos posted on online shopping malls – all of which are publicly available – are all considered works, so there was a dispute about whether they could be used for AI models without the permission of the copyright holder.

For more detailed information, please check Techm’s column at ​Main Contents of the Amended Copyright Act

Anil Moon, Senior Partner at Law Firm Veat, graduated from Yonsei University’s School of Engineering and Computer Science and received an award (Silver Prize) for 7th place in the Korean regional finals of the ACM-ICPC, the world’s largest and most prestigious programming competition for university students. He also worked as a Manager at Naver Corporation’s Search Development Center, researching document ranking systems (search engine ranking systems) and distributed databases. Furthermore, he is a member of the Korean Copyright Committee's Open Source Software License Expert Community and provides legal advice based on his technical understanding of software and intellectual property.

Thank you.