Japan Establishes Organization on AI Safety Evaluation

   Tokyo, Feb. 14 (Jiji Press)--The Japanese government established an organization on Wednesday to examine safety evaluation methods for generative artificial intelligence.

   The AI Safety Institute aims to establish evaluation methods to tackle risks such as the generation of disinformation in cooperation with related organizations in the United States and Britain.

   "We'll consider standards and methods for safety assessments to realize safe, secure and reliable AI," Chief Cabinet Secretary Yoshimasa Hayashi told a news conference.

   The AISI was established within the Information-Technology Promotion Agency, an independent administrative agency. The new entity will have several dozen employees, including from the Digital Agency and the private sector.

   The new entity is headed by AI researcher Akiko Murakami, chief digital officer at Sompo Japan Insurance Inc. who previously worked at IBM Japan Ltd.

   Japan and other Group of Seven industrialized nations are working on AI safety standards under their guidelines that call for AI developers to take risk reduction measures, such as making learning data more transparent before putting AI services on the market.
END