Japan AI Safety Institute

The Japan AI Safety Institute (J-AISI) was officially launched on February 14, 2024, following an announcement by Prime Minister Kishida on December 21, 2023. Its establishment reflected Japan's role in leading the Hiroshima AI Process and its commitment to responsible AI governance at the national and international level.

J-AISI is housed within the Information-technology Promotion Agency (IPA), a Japanese government body under the jurisdiction of the Ministry of Economy, Trade and Industry (METI) and the Digital Agency. Its secretariat consists of five teams with approximately 30 staff as of early 2025, including full-time and part-time personnel seconded from companies, ministries, and IPA itself.

The institute describes itself not as an R&D organization but as a hub for AI safety in Japan. Its core work spans six areas: developing safety evaluation frameworks and standards; conducting domestic outreach and education in partnership with academia and industry; coordinating internationally with peer institutions; conducting technical research on evaluation methodologies and testing tools; addressing AI security and cybersecurity intersections; and managing partnership agreements with international AI safety organizations.

J-AISI has published major outputs including the National Status Report on AI Safety in Japan 2024, a guide to evaluation perspectives on AI safety, healthcare AI safety evaluation guidelines, a Chief AI Officer Guidebook, and AI incident response resources. It participates in the international AI Safety Institute network and has conducted joint testing exercises with counterpart institutions. Advisors include Arisa Ema of Tokyo University and Riken, and Hideaki Kitano of Sony Group.