As debate grows over slowing AI development, speakers at a Seoul seminar said South Korea should raise the pace of responsibility through verification, control and recovery, rather than delay frontier AI. They called for Korea to design AI safety standards needed by both the United States and China, focusing on functional cooperation over value-based blocs. Participants also urged broader AI system evaluations beyond model scores and advocated Korea-specific benchmarks, risk-based management and stronger operational readiness in firms.

South Korea\'s Ministry of Science and ICT has moved its AI Policy Office into a second phase after appointing a new director. The office, created in a government reorganisation last year, oversees AI policy, industry promotion, computing infrastructure, data and safety and trust. The first phase focused on securing GPUs and advancing a homegrown AI foundation model project. The ministry now faces tasks including expanding evaluation standards, rolling out public-facing services and accelerating AI data centre development.