Law firm YulChon will participate with OSBC in the National IT Industry Promotion Agency (NIPA)'s support program for open-source artificial intelligence (AI) and software development. Their role is to check for copyright, license, personal information, and regulatory risks that can arise when using open-source software and external AI models in the AI development process.
YulChon said it will carry out open-source AI governance consulting with OSBC for 11 corporations selected as performers in NIPA's "open-source AI/SW development and utilization support program." Open-source AI governance refers to internal corporate procedures that systematically manage the open-source software, AI models, data, and license terms used in AI development.
This year's program, totaling 9.2 billion won, was launched for the first time to support development, utilization, disclosure, and diffusion of open-source AI and software led by domestic corporations.
Recently, as corporations rapidly adopt AI coding tools and open-source AI models, legal risks are also growing. If external libraries or open-source software that developers did not recognize are included in a product, license violations can occur. There is also an assessment that AI models, which combine data, code, and weights, blur the lines of copyright and usage rights more than general software.
YulChon and OSBC will review AI compliance alongside open-source governance in this consulting. AI compliance is a framework that ensures adherence to relevant laws and regulations and internal management standards when developing and operating AI.
In AI compliance, they will review corporation-specific obligations under the Basic AI Act, whether a system qualifies as high-impact AI, and advance notice and labeling obligations for Generative AI services. High-impact AI refers to AI that can significantly affect life, safety, and fundamental rights. They will also examine copyright issues for training data and AI outputs, as well as issues under the Personal Information Protection Act.
In open-source governance, they will check the license terms of the open-source AI models and software used by the performing corporations. Because open-source AI models can have different conditions for commercial use or distribution by version, it is important to identify in advance the risks that corporations may face at the service launch stage.
YulChon and OSBC also plan to build an "open-source AI profile" that reflects domestic laws and systems and global standards such as the European Union (EU) AI Act and SPDX 3.0. An AI profile is a document that organizes the models, data, software components, licenses, and risk factors included in an AI system. They plan to apply this to the 11 performing corporations to address on-site issues.
Im Hyung-ju, head of YulChon's AIDC Center, said, "This program is significant in that it builds an AI governance framework suited to domestic corporate environments and applies and validates it with actual corporations," and added, "We are now able to review complex legal issues—such as the Basic AI Act, the Copyright Act, and the EU AI Act—in connection with corporations' development and commercialization processes."
Kim Taek-wan, CEO of OSBC, said, "Corporations must manage not only the AI model itself but also the open source included in the model, training data, license terms, and dependencies on external services," and noted, "To ensure transparency and traceability of AI systems, it is important to establish a management framework based on open-source governance and AI profiles from the development stage."