Major nations scale up AI risk verification

Governments leading AI development have been strengthening systems to verify the risks of advanced AI models, building dedicated safety institutes and evaluation capacity. The push reflects growing attention to the safety testing of large language models (LLM) and other frontier AI systems as deployment accelerates.

Korea's evaluation capacity limited to 6 technical staff

South Korea's AI safety evaluation body, by comparison, is reported to have only six technical personnel in its evaluation office. The figure has circulated as evidence that Korea's AI safety infrastructure lags behind that of other major countries stepping up risk verification efforts.

A caveat on the number

The "6" refers specifically to technical staff within the evaluation office, not the organization's total workforce. This caveat makes it difficult to treat the headline figure as a complete picture of Korea's AI safety evaluation manpower, since the body may have additional personnel in other roles.

The bottom line

Korea does operate a small AI safety evaluation workforce relative to the scale of the task, and the six-person technical staffing figure is accurate as far as it goes. But because the number is limited to one category of staff, it should not be read as the country's total capacity for AI safety oversight.

The claim that Korea relies on only six AI safety evaluation staff is Mostly True: the figure is accurate for the evaluation office's technical personnel, but it does not represent the full extent of the country's AI safety workforce.