‘2026 International AI Safety Report’ Presentation and Discussion Event Held

On May 15, 2026, ‘2026 International AI Safety Report’ Presentation and Discussion event was hosted by the British Embassy Seoul and co-organized by the Korea Development Institute(KDI) and UK AI Security Institute(UK AISI). The event brought together leading domestic and international experts in AI safety to discuss the report’s key findings and their policy implications.

At the opening ceremony, Jack Minty, Counsellor for Technology & Digital Trade at the British Embassy Seoul, underscored that “deepened international cooperation is essential for effective AI risk management.” Dr. Jungwook Kim, the Executive Director of the Center for International Development(CID) at KDI, highlighted that “AI is rapidly transforming our daily lives and society, making cross-border collaboration and technological understanding more urgent than ever,” expressing his hope that the event would serve as an opportunity to deepen mutual understanding of AI safety and strengthen bilateral cooperation between Korea and the UK.

Stephan Clare, lead writer of the 2026 International AI Safety Report, then presented the latest developments in AI capabilities alongside the emerging risks they pose and potential responses. While AI systems have demonstrated remarkable performance in areas such as coding, mathematics, and translation, he noted that significant limitations remain in handling complex tasks and addressing hallucination problems. He emphasized that AI's rapid advancement is simultaneously giving rise to technical and social risks — including cyberattacks, system malfunctions, and labor market disruption — making the establishment of an evidence-based risk management framework an urgent priority.

Myuhng-Ju Kim, the Executive Director of the Korea AI Safety Institute, announced that a Korean edition of the report is currently in preparation, incorporating practical cases from Korea's industry, academia, and public sector. Dr. Kim explained that 26 domestic AI safety cases have been mapped onto the original report's framework, covering areas such as technical safeguards, risk management practices, misuse risks, and hallucination mitigation.

The event concluded with a panel discussion moderated by Dr. Jungwook Kim featuring Stephan Clare, Dr. Myuhng-Ju Kim, Professor Ahreum Hong from Kyung Hee University, and Professor Sangkyun Lee from Korea University. The panel discussion explored a range of topics including domestic and international AI risk trends, AI-related risks from a cyber security perspective, and the challenges of AI adoption in developing countries’ public sector. The discussion reaffirmed that as AI risks continue to evolve in increasingly concrete and complex ways, building effective risk response frameworks and strengthening international cooperation are indispensable.

Written by: Jieun Park, Research Associate, Division of Development Research, Center for International Development (CID), 044-550-4421, jieunp@kdi.re.kr