Personal Information Protection Commission Hosts Open Source Day… Pre-event for GPA General Assembly
(Seoul = Yonhap News) Reporter Minji Cha – An international forum was held to safely utilize open source (a method where source code is made public for anyone to use, modify, and redistribute freely) in the era of artificial intelligence (AI).
The Personal Information Protection Commission announced that it held the ‘Open Source Day’ event on the 15th at the Grand Hyatt Hotel in Seoul.
Global companies such as Google, Meta, Microsoft, OpenAI, Naver, Select Star, and AIM Intelligence, along with domestic researchers and overseas regulatory bodies, attended to discuss ways to establish a responsible open-source AI ecosystem.
This event was organized as a pre-event for the 47th Global Privacy Assembly (GPA), the world’s largest assembly of supervisory authorities, attended by 148 personal information supervisory bodies from 95 countries.
Google introduced its large language model (LLM) quality evaluation tool, prompt optimization functions, and safety enhancement measures using its platform ‘Vertex AI.’
AIM Intelligence, which advanced Meta’s Llama Guard to fit the Korean context, shared various security and safety challenges faced in the field.
Microsoft presented an agent AI construction case based on ‘Azure AI Foundry,’ while Naver introduced tools for safe open-source use, such as HyperCLOVA X, public datasets, and AI safety frameworks.
OpenAI introduced its newly released open-source models (gpt-oss-20b/120b), and Select Star presented its reliability verification solution DATUMO Eval as well as the case of establishing Korea’s first AI reliability evaluation benchmark.
In the latter part, privacy supervisory authorities from four countries, including Korea, the UK, Italy, and Brazil, held a roundtable.
The four national bodies agreed on the need for ongoing international discussions and cooperation to establish a safe AI ecosystem, anticipating an era of ‘agentic AI’ based on autonomy.
In a survey conducted by the Personal Information Protection Commission before the event, 62% of the 70 developers, researchers, and company officials surveyed said they had experience with adopting and utilizing open source. The proportion who considered safety during the fine-tuning process of using an open source model was also as high as 77%.
Choi Jang-hyuk, Vice-Chairman of the Personal Information Protection Commission, stated, “This Open Source Day is significant as the first public forum to jointly consider open source, a foundation for innovative services like agent AI, and personal information protection.” He added, “We will actively incorporate voices from the field into policy so that companies and researchers can confidently utilize open source.”
Kim Hwi-gang, a non-standing commissioner of the Personal Information Protection Commission, said, “The culture of openness and sharing in open source is driving the acceleration of technological dissemination and innovation across industries and society.” He expressed his expectation that it would become a starting point for open and trustworthy AI development.