A national honor for AI privacy work
Max Sun-woo, dean of Soongsil University’s AI College and head of its AI Graduate School, received the Hongjo Order of Service Merit, a South Korean government decoration recognizing contributions to public service. The award was presented at the sixth Personal Information Protection Day ceremony, held on September 30 at the El Tower in Seoul.
Sun-woo was recognized for helping organize ways to identify and reduce privacy threats linked to artificial intelligence. In particular, he served as chair of the risk-management division of a public-private policy council on AI privacy.
The award recognizes efforts to make AI privacy risks easier to assess and address—not only after a problem occurs, but as systems are designed and used.
The ceremony, hosted by South Korea’s Personal Information Protection Commission under the theme “An AI Innovation Society Advancing Through Trust,” brought together about 350 people working in privacy protection and related fields. Organizers presented 13 government awards and 52 commission chairperson commendations; Sun-woo was among the recipients of a government decoration.
How the AI privacy council works
The public-private council was launched by the Personal Information Protection Commission in February. Its 37 members come from industry, universities, the legal profession, and civic organizations, and work together on rules that consider both the benefits and risks of AI.
The council is divided into three areas:
- Data-handling standards: how personal information should be collected, used, and stored.
- Risk management: how to identify possible harms and reduce their likelihood or impact.
- Individual rights: how people can understand and exercise rights related to their personal information.
As head of the risk-management division, Sun-woo helped examine risks that can arise at different stages of personal-data processing. These risks can overlap: for example, an AI service might collect more information than necessary, infer sensitive details from seemingly ordinary data, or expose information through a security weakness.
Why privacy risks can be difficult to spot
AI systems may combine large amounts of information and use patterns within it to make predictions. A user might share a few routine details with an online service, for instance, while the system uses them alongside other data to infer something more personal. A risk framework can help teams ask what information is involved, who might be affected, and what safeguards should be put in place.
This work is intended to make privacy protection part of the development process rather than a last-minute check. That distinction matters as AI tools move into everyday settings such as customer support, education, health services, and workplace software.
Research and education at Soongsil University
Soongsil University’s AI College brings together four undergraduate areas: artificial intelligence, information security, software, and AI systems. Its AI Graduate School runs programs in AI convergence security and AI biotechnology, combining technical training with research in fields where AI and other disciplines meet.
The university’s AI Safety Research Center has also developed Muneme, a multi-agent vulnerability-analysis system. A multi-agent system uses several AI agents—software components that perform assigned tasks and share results—to examine a problem from different angles.
In the reported evaluation, Muneme achieved a 91.0% single-attempt success rate across 1,507 tasks in CyberGym Level 1, a practical benchmark (a standardized test used to compare system performance) for evaluating cybersecurity capabilities. A single-attempt rate measures how often the system succeeds on its first try, without relying on repeated attempts.
What vulnerability analysis means
A vulnerability is a weakness in software or a computer system that could be exploited to cause harm, such as unauthorized access or data theft. Finding these weaknesses early can help developers fix them before a product is widely used.
AI-based analysis can assist security teams by examining code and identifying potential problems. It does not replace human review: specialists still need to verify findings, judge how serious a weakness is, and decide how to correct it.
Sun-woo’s response and the wider challenge
Sun-woo said the recognition was especially meaningful because many people had worked over the years to advance privacy law and policy. He described the honor as one shared with others who have contributed to the field and thanked the commission and related organizations.
“In the AI era, technologies that can achieve both personal information protection and responsible use will become increasingly important. As a technology researcher, I will do my part to build a trustworthy AI society.”
The challenge he highlighted is central to AI governance: personal information can help services work better, but its use can also create risks for the people it describes. For example, a service may need some information to provide a personalized feature, yet collecting or retaining unnecessary details can increase the consequences of a leak or misuse.
Effective safeguards therefore need to address the entire process—from deciding what data is needed to controlling access, testing systems, and responding when something goes wrong. Clear rules and technical protections work best when they reinforce each other.
What the award may signal
The honor draws attention to the role researchers can play in connecting AI engineering with privacy policy. As AI services become more common, institutions and companies will face pressure to explain what information their systems use and how they limit potential harm.
- For developers: privacy and security checks can be built into design and testing rather than added after launch.
- For organizations: risk assessments can help identify weak points before personal information is exposed or misused.
- For the public: clear explanations and meaningful choices can make it easier to understand how personal data affects an AI service.
The award does not by itself establish that AI privacy risks have been solved. Instead, it recognizes work to create more systematic ways of identifying and reducing them—a task likely to grow in importance as AI tools become more capable and more deeply embedded in daily life.
Sun-woo’s current roles
Sun-woo serves as dean of Soongsil University’s AI College and head of its AI Graduate School. He also chairs the risk-management division of the public-private AI privacy policy council and leads the safety and trust division of the Agentic Alliance.
His research focuses on AI security and safety, including technologies designed to counter AI-related attacks. The award recognizes his contribution to privacy-risk work within that broader effort to develop AI systems that are safer and more trustworthy.
