About Center for Trustworthy AI
Our Research
Establishing and Materializing Trustworthy AI Standards
Our primary research objective is to create an AI trustworthiness assessment framework that aligns with our society and legal system. To achieve this, we propose the concept and evaluation criteria for Trustworthy AI, incorporating ethical and legal requirements, while researching and developing essential technologies for its verification and implementation.
Throughout this process, we strive to bridge the gap between technology and ethical/legal obligations through close interdisciplinary research. Our goal is to develop practical measures and solutions that developers can readily apply in the field, while simultaneously establishing a foundation for interdisciplinary convergent research and nurturing convergent talents.
Establishing and Materializing Trustworthy AI Standards
Implementation Plan
Phase 1 (2022-2024)
Our goal for the first phase is to develop the concept and evaluation criteria for Trustworthy AI that embody the ethical and legal requirements of society. We plan to publish technical documents outlining the standard procedures necessary for their verification and implementation. Following this, we will assess whether AI models or datasets align with the established concepts and evaluation criteria. Additionally, we will conduct research on core technologies to accurately rectify and ensure trustworthiness, with the results being published in academic papers.
Phase 2 (2025-2028)
During the second phase, we will build upon Phase 1 research on the verification framework and core technologies, with a particular focus on establishing a system to verify bias, harmfulness, identifiability, safety, and security - key challenges in generative models. In addition, we will develop software modules to ensure that our research findings are applicable for the industry. We also intend to actively create a comprehensive education program to nurture convergent talents.
Expected Outcomes
①
Academic and Technological
Advances in core technologies and interfaces in the field of trustworthy AI
Strengthening technological competitiveness through effective collaboration across various disciplines
Strengthening educational programs through developing curricula that incorporate research achievements
②
Economic and Industrial
Maximizing research impact through standardization and open-source development
Enhancing the usability of interfaces and frameworks by developing the necessary components for commercialization in various fields
Promoting research collaborations among universities, research institutes, and companies in Korea and overseas
③
Social
Nurturing interdisciplinary experts in the field of trustworthy AI
Enhancing the impact and social acceptability of AI technology
Improving the trustworthiness of Korean natural language processing and enriching its culture
